ABSTRACT. In this paper, an enhanced OFDM (E-OFDM) scheme, using uniform subband division, is proposed to improve the spectral efficiency. In the E-OFDM system, the divided subbands realize parallel transmission in the time domain, which decreases the sampling rate. Meanwhile, a multistage interpolator architecture is applied to each subband, resulting in a reduced computation complexity. Through subtly designing filters, the out-of-band emission of the signal is significantly depressed. Furthermore, narrow frequency guard intervals are reserved between subbands by filtering to resist the inter-subband interference, which can maximize the spectrum utilization. In the doubly selective fading channel, the Doppler and multipath diversity gains are retrieved from the oversampled signal. According to the two diversity gains, the  bit-error-rate performance of the proposed scheme gets lower for increasing the Doppler spread. Experimental results show that the spectral efficiency is increased up to about 98.2% for the digital terrestrial multimedia broadcasting (DTMB) and digital video broadcasting (DVB) systems.

ABSTRACT. Two-dimensional non-uniform constellations (2D-NUC) have been adopted for the most recent terrestrial broadcasting system called ATSC 3.0. They are known for being more efficient than one-dimensional non-uniform constellations (1D-NUC) and uniform constellations (UC). They can be used in any communication system with the bit-interleaved coded modulation (BICM) structure. However, one of the main challenges of such constellations is their desing for optimal behaviour in a wide range of cases and the design of a demapper that exploits their advantages. This paper presents two different condensation methodologies to design efficient 2D-NUCs simple in design and with limited demapping complexity. The proposal provides a complexity reduction of the design and demapping processes in the range from 13% to 94% .The demapping stage provides performance losses less than 0.1 dB if compared with standard 2D-NUC demapping.

ABSTRACT. In this paper, multiple algorithms to design non-uniform constellations (NUCs) are reviewed, it is found that the degrees of freedom (DOF) is still large when designing 2-dimensional NUC (2D-NUC). To this end, we propose a two-step algorithm to design 2D-NUC, which consists of initial constellations design and a subsequent iterative optimization. By imposing more limitations on the initial constellations, the number of parameters to be optimized is reduced. Based on that, a set of 2D-NUC which outperforms that adopted by ATSC 3.0 is finally presented by applying the proposed iterative optimization algorithm.

ABSTRACT. In this paper, we report on our study into the relationship between the image resolution and impressions such as “looks delicious” and/or “looks beautiful” through subjective evaluations, and then show how higher resolutions work in terms of enhancing a viewer’s impression of an image. We found that higher resolutions close to the resolution limit of the human eye affected the image impressions provided by familiar subjects. In addition, we found that the impression enhancement effects of higher resolutions were observed even when the viewing distance was beyond the standard one.

ABSTRACT. One of the goals of next generation broadcasting services is to provide realistic media contents to the users. Realistic media contents are media with multiple sensorial effects (mulsemedia) aimed at increasing user's experience through the five sense representation (i.e., taste, sight, touch, smell, and hearing). In this research domain, a key issue is the measure of the impact of mulsemedia on the perceived user's quality of experience (QoE). This paper provides an updated overview of the researches conducted in the area QoE for mulsemedia services. Results are summarized based on subjective quality assessments for audio-visual sequences which are enriched with effects, such as ambient light, wind and vibration effects. This paper aims at providing the broadcasting community with a survey to better understand the role that mulsemedia could play in the QoE enhancement in future broadcasting services. With this purpose, the research challenges that may encourage further research within this emerging domain are described.

ABSTRACT. We present an experimental study of the sensitivity of stereoscopic 3D-Video streams to lossy networks. By means of a controlled network testbed, we impaired and assessed the quality of a broad range of 3D stereoscopic video sequences considering different content types, compression levels and packet loss rates. Our study and methodology are meant to provide service providers with the means to pinpoint the working boundaries of their videosets in face of different network conditions.

ABSTRACT. Color change is an important factor in image quality assessment (IQA). Changing color manually is a hard work and improper process may destroy the perceptual quality. This problem has been largely ignored in traditional IQA which mainly focused on evaluating blur, noise and compression loss,etc. One bottleneck in this field is the lack of databases. Existing databases have very few distortion types related with color changes. In this paper, we construct a new database specially designed for color changed IQA problem. We then test a majority of state-of-the-art IQA methods on this database and prove that existing algorithms are indeed not suitable for color changed distortion. Therefore, we build a novel matrix considering features in color theory and statistics. The new model can effectively evaluate the color changed image quality.

ABSTRACT. Long Term Evolution Advanced (LTE-A) describes the latest evolution of the mobile network standard of the 3rd Generation Partnership Project (3GPP). Cellular wireless communication standards, such as LTE-A, are primarily developed for the content distribution of user specific data via unicast transmissions. The exponential growth in data traffic in mobile networks might require alternative approaches for content distribution in order to relief network cells from load congestion. The 3GPP evolved Multimedia Broadcast Multicast Service (eMBMS) and Single-cell Point-to-Multipoint (SC-PTM) enable multicast/broadcast transmissions of content as an alternative transmission layer to the unicast transmission layer. SC-PTM and eMBMS rely on the existing Low Tower, Low Power (LTLP) infrastructure of current mobile access networks, effectively reducing the number of resources available for unicast services. The Tower Overlay over LTE-Advanced+ (TOoL+) system, developed by Technische Universitaet Braunschweig, introduces an additional transmission layer by offloading popular content to a dedicated broadcast carrier transmitted from a High Tower, High Power (HTHP) infrastructure. The required modifications of the LTE-A standard are indicated by the term ’LTE-A+’. To evaluate the capabilities of the TOoL+ system in a realistic environment, we integrated the 3GPP broadcasting standards and TOoL+ into the Simulator for Mobile Networks (SiMoNe) environment, a software platform for realistic mobile network simulations developed by Technische Universitaet Braunschweig. In this paper we present the implementation details and the impact of the different broadcasting layers on the cell load in a realistic network scenario.

ABSTRACT. 3GPP is currently studying the improvement on the LTE eMBMS system to provide more capabilities and flexibilties on delivering multicast/broadcast services using the LTE infrastructure. This paper investigates the performance of an improved LTE eMBMS system using layered-division-multiplexing (LDM) technology and longer cyclic prefix (CP). The link-layer performance of using two-layer LDM in eMBMS system is first investigated with extensive computer simulations. Coverage analysis is then conducted on future eMBMS services with different cyclic prefix (CP) lengths, to allow more efficient deployment of multicast/broadcast single-frequency-network (MBSFN) for different cell sizes. The coverage performance is then used to demonstrate the capability of the improved eMBMS system with LDM to simultaneously deliver high-definition (HD) indoor/mobile services and ultra-HD fixed services targeting receivers with rooftop antennas

ABSTRACT. The demand for video services (i.e., video streaming, video downloading, video conferences, concert videos, etc.) in mobile networks is rapidly increasing. Video will account for more than 69% of mobile data traffic by 2018 [1]. For this reason group-oriented and on demand services will play a key role in the future wireless systems, posing new challenges in the design of techniques to improve the throughput and the delays to provide those services. Moreover, the scalability for large amount of users every time and everywhere must be guaranteed. 3rd Generation Partnership Project (3GPP) introduced evolved Multimedia Broadcast and Multicast services (MBMS), to allow Long Term Evolution (LTE) to provide data transmissions from a single source to multiple devices [2]. MBMS utilizes a common channel to send the same data to multiple receivers, thereby minimizing the utilization of network resources. Efficient link adaptation is a key feature that allows LTE systems to react to variations of the channel quality. Amongst all the options possible in LTE, the division of potential receivers in different groups with different available SNR values is a well-accepted technique for enabling a more efficient exploitation of the spectral resources. This technique is known as multicast grouping (MG) [3]. Each user in a MG transmits Channel Quality Indicator (CQI) feedbacks to the base station (BS), which decides on a per-group basis the number of Resource Blocks (RBs) to allocate the modulation and coding schemes (MCSs). Single cell point to multipoint (SC-PTM) mechanism [4] uses the MBMS for improving radio efficiency and reduce latency, allowing efficient radio utilization and flexible deployment of a number of applications and services. The broadcast/multicast area can be dynamically adjusted cell by cell according to user’s distribution, interests, and movement. The main objective of this work is to compare the benefits and disadvantages of three different approaches proposed for managing multicast services

ABSTRACT. Low Power Wide Area Networks (LPWANs) are a network type for connecting thousands of ultra-low power sensor-nodes to a base-station network, mainly using license-exempt frequency bands. Although the sensor-nodes use a transmit power of typically 10mW, such systems can reach a transmit distance of up to 40km by means of ultra-low payload bit-rates. However, in case of bi-directional communication the link budget between the uplink and downlink is highly imbalanced with the problem being on the downlink. Common solutions such as increasing the transmit power on the downlink are not possible, as the maximum transmit power and duty cycle in the downlink are limited by regulations. The use of broadcast signals for the downlink, e.g. as part of DVB-T2 Future Extension Frames, may solve these problems. Therefore, a combination of LPWANs and broadcast systems seams very appealing from a technical point of view.

ABSTRACT. Channel estimation for millimeter wave (mmWave) systems over time-varying channels is a challenging problem, since a large number of channel coefficients need to be estimated. In this paper, by exploiting the sparsity of mmWave channel in the angular domain, we propose an efficient sparse channel estimation scheme based on compressive sensing (CS) theory. Specifically, considering that the angles of arrivals/departures (AoAs/AoDs) vary more slowly than the path gains, we formulate the channel estimation into a block-sparse signal recovery problem, and then propose a novel greedy algorithm consistent with this structure to estimate AoAs/AoDs. Based on the estimated angles, we design the optimal training hybrid precoders and combiners to maximize array gains, followed by estimating the path gains utilizing the least square (LS) method. The simulation results demonstrate that our proposed scheme performs better than the existing mmWave channel estimators in both estimation accuracy and spectral efficiency over time-invariant channels, and further verify that our proposed scheme is suitable for timevarying channels.

ABSTRACT. Extended Abstract

I. Introduction Single-frequency-network (SFN) is a preferred deployment solution for digital TV (DTV) broadcasting systems. By emitting the same broadcasting services from all transmitters using the same channel, using SFN can achieve much higher spectrum efficiency when delivering broadcasting services over a large geographical area. It was further shown in [1] that using SFN can also bring significant transmission power saving than the current single-transmitter solution. In an SFN, multiple transmitters send out the same signals that are frequency-locked and time-aligned. This provides much higher chance for the receivers to constructly combine the received signals from multiple adjacent transmitters. However, the nature of SFN does not provide a mechanism to deliver distributed local services that are specific to each transmitter, also called local contents. Delivering local contents has always been an attractive application scenario for broadcasters, which can be used for local news, location-based applications, as well as local advertisement. Therefore, implementing local content insertion (LCI) in DTV SFNs becomes a highly interested issue. In literature, different mechanisms have been proposed for LCI, based on hierarchical modulation [2], orthogonal time-division-multiplexing (TDM) or frequency-division-multiplexing (FDM) [3], and multiple-input-multiple-output (MIMO) technologies [4]. While using hierarchical modulation results in low-complexity receivers, the throughput of the local services is very limited. Using TDM/FDM to deliver local contents needs a network-wide co-ordication on the time/frequency allocation in order to avoid adjacent-cell interference (ACI), which could results in low frequency utilization when not all transmitters are required to deliver local contents. Using MIMO requires additional complexity at the receivers and requires good signal to noise ratio (SNR) conditions to provide good transmission capacity. In addition, these traditional methods are likely to suffer from service coverage gaps for the local services between adjacent transmitters so as to avoid the adjacent cell interference (ACI). The recently developed next generation DTV system (a.k.a., the ATSC 3.0 system), adopted a new non-orthogonal physical-layer multiplexing technology to efficiently deliver multiple broadcasting services with different quality of service (QoS) requirements, which is called the layered-division-multiplexing (LDM) technology. In [5], it was shown that the LDM technology offers a new option to deliver local services within an ATSC 3.0 SFN environment. The basic idea of LDM is to use a multi-layer signal transmission structure to deliver different types of services, where all the signal layers occupy the same time and frequency resource. The multiple layers are allocated with different transmission powers and signal configurations, determined by the QoS of the targeted serrvices. The ATSC 3.0 adopted a two-layer LDM structure, where the signal layer with higher-power layer is called the Core Layer (CL) and the lower-power layer is called the Enhanced Layer (EL). The transmitter and receiver block diagrams of a two-layer ATSC 3.0 system configured for local content insertion are shown in Figure 1. The CL is configured to deliver a global high-throughput UHDTV service, while the EL is configured to deliver local services within each transmitter’s coverage area. In this scenario, both CL and EL services are designed for fixed receivers with powerful antennas (e.g., rooftop antennas). To decode the local services, the receivers need to decode the global services first and perform successive signal cancellation (SSC). It was shown in [5] that using the LDM can deliver local contents at any time, any location (from any transmitters), and over a flexible coverage areas. These advantages are very difficult to achieve with the traditional solutions. It is further shown in [5] that, over a typical 6 MHz TV channel, a properly configured two-layer LDM delivers an UHDTV service in the CL and local services of 720p-HD-quality in the EL.

Figure 1 : Local Content Insertion using Two-Layer LDM II. LDM-EL Local Service Detection in SFN So far, the performances of local service delivery using LDM were obtained by computer simulations with the assumption of perfect channel estimation at the receivers. These performances serve as the upper bounds of the achievable local service throughput for practical situations. Let’s consider the local service insertion in an SFN network with two transmitters, as shown in Figure 2, where the LDM-CL is used to deliver global services and the LDM-EL is used for local services. The received signal can be expressed as, (1) where the sCL contains the global service, sEL-1 and sEL-2 contain the local services emitted by the two transmitters, H1 and H2 are the channel responses from the two transmitters to the receiver, n(k) is the AWGN.

Figure 2 : SFN Scenario of Local Content Insertion To decode the local service contained in sEL-1, the receiver first needs to cancel the CL signal. For an ATSC 3.0 SFN, the two adjacent cells transmit signals with the same pilot pattern. Therefore, a regular channel estimator will generate an estimate of the combined channel for the CL signal, H, in (1). The decision symbols of the sEL-1 is now calculated as, (2) where ICL is the residual CL signal after SSC. It becomes obvious from (2) that decoding the sEL-1 requires the knowledge of the channel response from the first transmitter (Tx1) to the receiver. The current ATSC 3.0 standard does not provide a mechanism for a receiver to achieve separate channel estimations when receiving multiple signals from different transmitters. Therefore, good channel estimation solutions need to be identified to achieve optimal detection performance on local service detection. III. Performance of Local Content Detection with Regular Channel Estimator In this section, the performance of local content detection for a receiver with normal channel estimator is evaluated by computer simulations. Let’s assuming the receiver is located in the area of Tx1. The regular channel estimator uses the in-band pilot and obtains an estimate of the combined channel response, H = H1 + H2. Using this channel estimate to decode sEL-1 will results in performance loss as comparing to the optimal detection uisng the actual channel estimate on H1. When the power of the received signal from Tx1 is significantly higher than that from Tx2, the combined channel, H, is very close to the first local channel, H1. In this case, using the conventional channel estimator can approach the optimal detection performance of the local service from Tx1. The performance loss starts to occur when the signal from Tx2 is has a power close to that from Tx1. In this case, the combined channel, H, is significantly different from the desired channel, H1. The final version of this paper will present: • the local service detection performance loss versus the relative power ratio of the two received signals from Tx1 and Tx2 are evaluated by computer simulations; • the performance loss characterized by the lower throughput of the local services, or higher system SNR required to detect the local services.

IV. Channel Estimation Strategies for Local Content Detection This section will propose several solutions to achieve more accurate channel estimation for better local content detection in ATSC 3.0 systems. The final version of this paper will discuss the following options: • Using ATSC 3.0 bootstrap signal to achieve separate channel estimate [6]; • Using ATSC 3.0 transmitter identification (TxID) signal to achieve separate channel estimate [7]; • Using directional antenna at the receiver to improve the channel estimation accuracy for local content signal detection;

References [1] L. Zhang, Y. Wu, K. Walker, W. Li, K. Salehian and A. Florea, “Improving LTE eMBMS with Extended OFDM Parameters and Layered-Division-Multiplexing,” IEEE Trans. Broadcasting, vol. 63, no. 1, Part I, March 2017. [2] J. Lopez-Sanchez, J. Zollner, S. Atungsiri, E. Stare, D. Gomez-Barquero, “Technical Solution for Local Service Insertion in DVB-NGH Signal Frequency Networks,” IEEE Trans. Broadcasting, vol. 60, no. 2, Part II, pp.293-301, June 2014. [3] D. Gomez-Barquero, C. Douillard, P. Moss, and V. Mignone, “DVB-NGH: The Next Generation of Digital Broadcast Service to Handheld Devices,” IEEE Trans. Broadcasting, vol. 60, no. 2, Part II, pp.246-257, June 2014. [4] Sungho Jeon, Jae-Shin Han, Robin Shrestha, Sung Ik Park, Ha-KyunMok, Heung Mook Kim and Jong-Soo Seo, “MIMO cloud Transmission Based on BICM-ID for High Data rate Local Content Delivery,” IEEE Trans. Broadcasting, vol. 61, no. 4, pp.580-589, Dec. 2015 [5] W. Li, Y. Wu, L. Zhang, K. Salehian, S. Laflèche, D. He, Y. Wang, Y. Guan, W. Zhang, J. Montalban, P. Angueira, M. Velez, S.-I. Park, J.-Y. Lee, and H.-M. Kim, “Using LDM to Achieve Seamless Local Service Insertion and Local Program Coverage in SFN Environment,” IEEE Trans. Broadcasting, online published, Jan. 2017. [6] Advanced Television Systems Committee (ATSC), “ATSC Standard: System Discovery and Signaling,” Doc. S321:2016, 23 March 2016. [7] S.-I. Park, W. Li, J.-Y. Lee, Y. Wu, X. Wang, S. Kwon, B.-M. Lim, H. M. Kim, N. Hur, L. Zhang, and J. Kim, “ATSC 3.0 Transmitter Identification Signals and Applications,” IEEE Trans. Broadcasting, online available, Jan. 2017.

ABSTRACT. To enable transmissions of large-volume content such as Super High-Vision, we are conducting research on digital terrestrial broadcasting and considering using low density parity check (LDPC) code as the forward error correction (FEC). For terrestrial broadcasting, the performance of forward error correction should be evaluated not only in AWGN environments but also in multipath environments. Moreover, it would be desirable for the new FEC to have better performance than conventional one even in multipath environments. Previous research evaluated the transmission performance of LDPC code by using computer simulations of a propagation model. The actual propagation characteristics were not evaluated. In this paper, we describe the transmission performance of convolution coded OFDM (CC-OFDM) and LDPC-coded OFDM (LDPC-OFDM) by computer simulation using the same signal structure and propagation channels, as measured in an outdoor experiment conducted in an urban area. We show that LDPC code has better transmission performance compared with CC in multipath environments and show how the transmission performance depends on the code rate and code length.

ABSTRACT. The frame of ATSC 3.0 consists of bootstrap, preamble and subframe in order, and the bootstrap signal can be used for initial synchronizations at the receiver. Also, it contains the signaling information needed for decoding of the preamble. Therefore, the bootstrap signals have to be reliably decoded even under a poor condition. In this paper, we derive the maximum-likelihood (ML) decision rule for the detection of the bootstrap signals and propose an improved decoding algorithm of bootstrap signals for ATSC 3.0 systems. The proposed algorithm improves the reliability of the channel estimate by averaging the channel estimates for previous symbols. The simulation results show that the proposed algorithm can improve symbol error rate (SER) performance by enhancing the reliability of the channel estimates.

ABSTRACT. Video streaming over networks has grown rapidly in recent years. Increasing focus has gradually turned from Quality of Service (QoS) awareness to user Quality of Experience (QoE) awareness. In this paper, we propose a unified QoE prediction framework for HEVC encoded video streaming. Firstly, the impact of network impairment on QoE is modelled by packet loss rate. Secondly, features of HEVC encoding process are analyzed and the most relevant one is selected to formulate the impact of HEVC encoder. Thirdly, the impact of video content is incorporated by analyzing the average number of bits for one coding unit (CU). Finally, the impacts of network impairment, HEVC encoder and video content are taken into account as multiplicative factors to develop a QoE metric. Experiments has been done for typical video sequences in terms of Pearson correlation coefficient, Spearman correlation coefficient and outlier ratio. Results show that the proposed QoE framework not only provides good correlation (93.4%) with subjective ratings, but also outperforms the existing results.

ABSTRACT. Multimedia streaming to mobile devices is one of the main sources of network congestion. As bandwidth requirements are continually increasing and users are becoming more quality-aware, there is a growing need for QoE-aware multimedia adaptations solutions. This paper presents a novel mechanism for automatic computation of quality threshold values up to which the video resolution can be decreased while still maintaining a predefined QoE level. The mechanism uses full-reference objective VQA metrics and rules for mapping their values to the subjective MOS scale. The results of a subjective study show that mapping rules for VQA metrics such as PSNR, SSIM and VIFp provide up to 72.22% exact MOS level match accuracy across different categories of multimedia clips. Moreover, accurate resolution threshold computation requires careful selection of the mapping rule to balance the under and overestimation of subjective video quality.

ABSTRACT. While traditional multimedia has primarily focused on vision and hearing, a new form of media delivery known as multiple sensorial media” (Mulsemedia) is actively growing and developing. Mulsemedia represents a revolutionary way of delivering multimedia by integrating additional components capable to reproduce other senses (e.g. touch, olfactory, gustatory) into multimedia content. In the context of classic multimedia applications, user quality of experience (QoE) has been extensively studied. Some QoE related research has also been performed in case of 2D video based Mulsemedia content, demonstrating the capability of Mulsemedia to increase user QoE as compared to classic multimedia. However, there is a need for further studies on the user QoE when experiencing Mulsemedia. This paper performs an extensive subjective study in order to investigate for the first time in the literature user QoE when subject to 3D video based Mulsemedia. The results obtained in case of the tests run for the 3D video based Mulsemedia are discussed and analysed in comparison with the results obtained for the same tests performed in 2D version. The user QoE associated to 3D video based Mulsemedia is demonstrated to be increased as compared to the user QoE associated to 2D video based experience.

ABSTRACT. Stereoscopic 3D (S3D) image technology has been extensively developed in the last decades. Visual discomfort such as eye strain, headache, fatigue, asthenopia, and other phenomena leading to a less pleasant viewing experience is still a potential issue in S3D applications. How to evaluate S3D image quality that related to visual discomfort is still a challenging problem. A larger number of studies have been done on S3D Image Quality Assessment (S3D IQA) where the subjective assessed S3D Image Databases play an important role. The subjective scores were collected for each S3D image in database with a number of viewers. Usually, Likert scale is adopted for observers to mark their subjective quality score, and then mean opinion score (MOS) is estimated. Due to the law of comparative judgment, the quality of subjective scores varies among observers and relates to the judgment method. This paper studied the quality of two subjective assessment methodologies - single stimulus (SS) and pairwise comparison (PC). Considering the S3D IQA as a S3D images’ quality ranking problem, we applied single stimulus and pairwise comparison subjective testing on a set of S3D images with known geometric distortions. From SS subjective testing results, the S3D images’ ranking can be derived by sorting MOSs directly. From PC subjective testing results, the ranking can be derived from DMOS scores. The distorted S3D images can be ranked via their geometric distortion parameters. The quality of subjective assessed results from SS and PC is then evaluated on the correlation between their ranking results to corresponding geometric distortions. With the collected MOSs for geometric distorted S3D image database, a deep-learning based S3D IQA model was used to study the model performance with the quality of subjective assessment methods.

ABSTRACT. I. INTRODUCTION Recently the mobile data traffic reached two major milestones and continues to increase and diversify. First milestone refers to the offload traffic which exceeded the cellular traffic for the first time in 2015. Globally, the total public number of Wi-Fi hotspots is expected to grow seven fold from 64.2 million in 2015 to 432.5 million by 2020. Originally, due to the high availability of Wi-Fi hotspots globally, and continued pressure on users imposed by data caps, the mobile users have opted for Wi-Fi wherever possible. Lately, technology advances including dual mode devices have enabled smarter switching between Wi-Fi and cellular networks. The second milestone reached was that the 4G traffic surpassed the 3G traffic. Although 4G connections represented only 14 percent of mobile connections in 2015, they have accounted for 47 percent of mobile data traffic, while 3G connections represented 34 percent of mobile connections, but carried 43 percent of the traffic [1]. This trend continues today and is expected to include 5G solutions whenever they will become available. A major contributor to this traffic growth is the enhancement process of mobile devices which resulted in user access to smartphones and tablets with high processing power, diverse screen sizes and very ever-greater high display resolutions. Moreover, video-on-demand services such as Vimeo [2], YouTube [3], Netflix [4] and Amazon Instant Video [5] offer diverse high data rate content which puts pressure in terms of resources on both network for delivery and device for processing display. In particular, mobile device innovations are being introduced at unprecedented speeds and have improved exponentially with time. The last years have seen a 20 times increase in cellular speed, a 30 times increase in Wi-Fi bandwidth and a 34 times increase in video resolution (compared to 2010). On the other hand, battery performance is not keeping up and battery capacity is estimated to increase two times only from 2010 to 2020, meaning that the battery performance only increases with 5% each year. Therefore, it is no wonder the battery performance ranks as the least satisfying aspect on smartphones and customer complaint calls about battery power have quadrupled since 2008. Finally, 71% of mobile device users around the world said the best way to improve devices would be to provide longer battery life [6]. Whereby, battery life has significant impact on mobile devices’ practical use, especially for multimedia streaming and video-on-demand applications. This fact motivates our work with the final goal to solve a double problem. First, traffic overload in high dense heterogeneous networks and heterogeneous devices. Secondly, energy saving in order to increase the battery life while maintaining high QoS level for video services. This paper proposes an energy-aware device-oriented video delivery algorithm (E-ARMANS) which will be integrated into the Adaptive Real-time Multi-user Access Network Selection (ARMANS) [7] framework in order to balance energy saving and traffic load balancing in heterogeneous network environments. II. RELATED WORKS Some studies were conducted for energy saving and battery consumption. The work in [8] proposed a cross-layer Battery and Stream-Aware Adaptive Multimedia Delivery mechanism (BaSe-AMy) that monitors energy consumption at the Wireless Network Card (WNIC) in the physical layer, packet loss rate at transport layer and stream playback duration in the application layer, and performs video streaming adaptation in order to increase battery life, while maintaining a high QoS. In [9] Trestian et al. presented an Energy–Quality–Cost Tradeoff in a Multimedia-Based Heterogeneous Wireless Network Environment with a model that takes into consideration the estimated energy consumption of the mobile device when running real-time applications, the monetary cost of the network, application requirements, and estimated network conditions in terms of average throughput. The user has a choice of three different networks: WLAN A, WLAN B, or the UMTS and five scenarios were considered: user near AP without background load, user far AP without background load, user near AP with background load, user far AP with background load, and a cellular scenario user near AP without background load. In [10] Zou et al. propose proposes a QoE-aware Energy-saving Device-Oriented Adaptive Scheme (E2DOAS) for mobile multimedia delivery over wireless networks. E2DOAS is a solution that optimizes the trade-off between the end-user perceived quality of the multimedia delivery and the mobile device energy saving. This trade-off is used in order to adapt the multimedia application delivery to the underlying wireless network conditions and multi-device characteristics. Making use of the device heterogeneity, E2DOAS balances the QoE and energy saving based on different end-user requirements, and adapts the multimedia streams according to the network conditions. The evaluation results show that E2DOAS finds the optimal trade-off between QoE and energy savings, outperforming three other schemes considered from the literature, in terms of estimated perceptual video quality, energy saving factor, average throughput and packet loss ratio. Moreover, in [11] the authors detail how the wireless environment impact the energy consumption of a mobile device while performing video on demand (VoD). The tests were conducted on an Android Mobile Device in a controlled wireless (IEEE 802.11g) environment in order to better understand the impact of each parameter on the energy consumption. In these above-mentioned works, energy saving techniques were applied in a wireless or heterogeneous context. However, architectures with overloaded heterogeneous networks, heterogeneous devices and at the same time heterogeneous type of traffic are actually fully or partially missing. E-ARMANS considers four type of traffic and the associated QoS, mobile devices making video requests characterized by four different screen resolutions and returns the best candidate network with the lowest energy consumption. III. ENERGY-AWARE ARMANS E-ARMANS monitors the energy consumption at the LTE/Wi-Fi Wireless Network Interface Card (WNIC) at physical layer, packet loss rate at transport layer and stream playback duration at application layer. The framework of the proposed solution is illustrated in Figure 1. The E-ARMANS scheduling framework consists of three main blocks: server-side, client-side and Mobile Network Operator (MNO)-side block. At the mobile client-side, both Wi-Fi and LTE compatible multi-radio devices characterize the user equipment (UE). Each UE could be connected to either an 802.11 access point (AP) or an Evolved Node B (eNodeB). The UE integrates several functional blocks as follows. The Network Discovery block provides the user a list of available wireless connectivity options based on their coverage footprint. The Device Characteristics block details information about screen resolution and real time battery level. The QoS Monitor block sends to the server the quality of the received services in terms of Peak Signal to Noise Ratio (PSNR), packet delay and packet loss rate (PLR). The Application Running and Stream Playback Duration block informs the E-ARMANS server about the type of traffic involved in the current user session, and QoS levels associated to each traffic type. The user type-ID and Device Characteristics block is employed to differentiate the user in terms of their device screen resolution and battery level. Four Wi-Fi classes (i.e., VO – voice, VI – audio-video, BE – best-effort, BK – background) [13] and four QoS Class Identifier (QCI) [10] for LTE have been considered. Information about the application running and the effective or predicted stream playback duration involved in a particular session is periodically sent at every Transmission Time Interval (TTI). Finally, UEs communicate to E-ARMANS server their position and speed in order to assign the best network when moving through different networks. The MNO-side integrates the real-time network characteristics transmitted by the network operator to the E-ARMANS server. Both LTE (i.e., eNodeB) and Wi-Fi (i.e., AP) transmitters are geolocalised and communicate their position and respective network IDs (i.e., eNodeB-ID and AP-ID, respectively). Each transmitter node also communicates the total amount of traffic load and corresponding distribution within different type of traffic. The E-ARMANS server-side gathers information from both mobile client-side and MNO-side. The E-ARMANS algorithm located at the E-ARMANS server-side will compute a network-ranking list based on the calculated quality of the multimedia stream. Always best connected devices is the core of the proposed mechanism either according to the traffic load and load balancing MNO-side, energy consumption, and the application type running mobile user-side. IV. IN SUMMARY This paper introduces a novel E-ARMANS algorithm combining load balancing and energy saving to obtain the maximum user satisfaction when video requests occur from heterogeneous devices. The proposed method allows increasing mobile device battery life and traffic load balancing in heterogeneous network environments. E-ARMANS has the following features: - considers several type of traffic (i.e., voice, video, best-effort, background) and their associated QoS; - defines a procedure for load balancing over heterogeneous network; - considers the type of user, mobility, device screen resolution and residual battery level; - considers a heterogeneous wireless network where LTE and 802.11n standards operate; - proposes a trade-off between energy-saving and load balancing such as good QoE levels are obtained. E-ARMANS has been compared with a no energy-aware algorithm in which energy optimization was not considered. Preliminary results show an increased battery life of 4%, 7%, 12% and 16% for 720p, 1080p, 1440p and 2160p screen resolution video delivery, respectively. REFERENCES [1] “Major Mobile Milestones – The Last 15 Years, and the Next Five,” blogs@Cisco - Cisco Blogs. [Online]. Available: http://blogs.cisco.com/sp/mobile-vni-major-mobile-milestones-the-last-15-years-and-the-next-five. [2] “Vimeo: Watch, upload, and share HD and 4K videos with no ads.” [Online]. Available: https://vimeo.com. [3] “YouTube.” [Online]. Available: https://www.youtube.com/. [4] “Can I stream Netflix in Ultra HD?,” Help Center. [Online]. Available: https://help.netflix.com/en/node/13444. [5] “Amazon.com: Amazon Video: Amazon Video.” [Online]. Available: https://www.amazon.com/Amazon-Video/b/ref=topnav_storetab_atv/166-1865049-8673925?_encoding=UTF8&node=2858778011. [6] “[Infographic] Resolving the Battery Challenge in Mobile Devices,” Voices@SamsungSemiconductor, 25-Feb-2013. . [7] M. Anedda, G.-M. Muntean, and M. Murroni, “Adaptive Real-time Multi-user Access Network Selection Algorithm for Load-balancing over Heterogeneous Wireless Networks,” in 2016 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), Nara, 2016, pp. 1-4. doi: 10.1109/BMSB.2016.7521907, 2016. [8] M. Kennedy, H. Venkataraman, and G. M. Muntean, “Battery and Stream-Aware Adaptive Multimedia Delivery for wireless devices,” in IEEE Local Computer Network Conference, 2010, pp. 843–846. [9] R. Trestian, O. Ormond, and G. M. Muntean, “Energy #x2013;Quality #x2013;Cost Tradeoff in a Multimedia-Based Heterogeneous Wireless Network Environment,” IEEE Trans. Broadcast., vol. 59, no. 2, pp. 340–357, Jun. 2013. [10] L. Zou, R. Trestian, and G.-M. Muntean, “E2DOAS: User experience meets energy saving for multi-device adaptive video delivery,” in 2015 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), 2015, pp. 444–449. [11] R. Trestian, A.-N. Moldovan, O. Ormond, and G.-M. Muntean, “Energy consumption analysis of video streaming to Android mobile devices,” 2012, pp. 444–452. [12] “ITU-R Recommendation BT.500-13 - Methodology for the subjective assessment of the quality of television pictures.” . [13] “IEEE SA - 802.11-2012 - IEEE Standard for Information technology--Telecommunications and information exchange between systems Local and metropolitan area networks--Specific requirements Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications.” [Online]. Available: https://standards.ieee.org/findstds/standard/802.11-2012.html. [14] K. Mahmud, M. Inoue, H. Murakami, M. Hasegawa, and H. Morikawa, “Measurement and usage of power consumption parameters of wireless interfaces in energy-aware multi-service mobile terminals,” in 15th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, 2004. PIMRC 2004, 2004, vol. 2, p. 1090–1094 Vol.2.

ABSTRACT. Consider AF cooperative networks with multiple radio frequency energy-harvesting (RFEH) relays. At each time instance, one relay is selected to forward the source message, while the remaining relays harvest RF energy of the source signal. In this work, we investigate power allocation and relay selection strategies to prolong the network lifetime since relays are assumed battery-powered with limited battery capacity. The relay selection strategies aim to minimize the loss of energy dissipation, reducing wasted energy, and increase the average harvested energy, in order to prolong the network lifetime. Compares with existing relay selection strategies which take channel state information (CSI), residual energy information(REI) into consideration, we propose a novel relay selection scheme based on opportunity cost of each choice by jointly considering CSI, REI and harvested energy information(HEI). It shows through simulation that the proposed scheme provides significant gain on the network lifetime.

ABSTRACT. The increasing demand of wireless telecommunication services is stressing the availability of radio spectrum to accommodate new services. Nevertheless, spectrum usage surveys performed in different regions reveal an under-utilization of radio spectrum. TV White Space (TVWS) technologies allow a dynamic usage of the Very High Frequency (VHF) and Ultra High Frequency (UHF) television channels. In this paper, we investigate the coverage range, capacity, and energy efficiency of IEEE 802.22(b) technology. We consider Ghent, Belgium, to evaluate a realistic scenario. The higher coverage range is 7.9 km for IEEE 802.22b, when using Modulation and Coding Scheme (MCS) 1/2 QPSK. The best trade-off between coverage and capacity in Ghent suburban scenario is achieved for the MCS 2/3 16 QAM, with an energy efficiency of 50 square km∙Mbps/W, a Base Station (BS) maximum delivered bitrate of 16.064 Mbps and a coverage range of 2.9 km. To ensure a percentage of users covered higher than 97% are required 20 BS.

ABSTRACT. Due to the increasing spectrum scarcity caused by data-consuming multimedia applications over mobile and smart devices, efficient utilization of radio spectrum has been receiving tremendous interest during the past decades. According to Cisco's networking visual index report, the mobile data traffic is expected to grow at a compound annual rate of 61% by 2018, an 11-fold increase over 2013, where mobile video traffic accounted for 55% of overall traffic. Given the dramatic growth of mobile traffic and increasing scarcity of spectral resources, efficient spectrum management, and sharing mechanisms becomes essential for future wireless networks.

There have been lots of related studies in addressing the spectrum shortage problem in the last twenty years. One direction of such efforts is to allocate new spectrum resources, i.e., exploring idle spectrum particularly in the millimeter wave (mmW) band. However, it is extremely difficult to provide wide coverage in mmW band due to the poor signal propagation characteristics at extremely high frequencies. Another direction is to improve the spectrum utilization rate through spectrum sharing among co-existing wireless networks using dynamic spectrum sharing or cognitive radio (CR) technology. As some frequency bands in the existing spectrum allocation are underutilized, for instance, UHF/VHF bands reserved for broadcasting, it is beneficial to explore the spectrum white spaces for mobile data communications due to the desirable propagation characteristics to support coverage and mobility. As a result, CR was first proposed about twenty years ago to provide secondary access to the unused spectrum while reducing the possible interference among the multitude of wireless systems through spectrum sensing. However, due to the limited sensing capability of devices and lack of timely information exchange between coexisting primary and secondary networks, it is extremely difficult to avoid misdetection and performance degradation of the primary users. Due to these risks, primary networks are reluctant to open their spectrum resources for sharing with secondary networks. The low industry interest on standardization or production of CR technology in the industry also demonstrates this realistic concern.

All in all, conventional spectrum sharing schemes, particularly those using CR techniques, have to rely on limited input and link level decision-making without coordination across co-existing networks. These challenges are further compounded by the fact that existing broadcasting and broadband wireless networks have a rigid architecture relying on vendor-specific configuration interfaces, and the operators only have indirect control of the network’s operation and resources. There is hardly or no effective timely information sharing among the independent, heterogeneous networks for the real-time spectrum access. Thanks to the ongoing development of the next-generation ATSC 3.0 DTV broadcast system with return channel, the convergence of broadcasting and broadband wireless systems becomes possible and beneficial, as the return link enables information collection through real-time updates from receivers and traditional approach of exclusive usage has become too costly for both systems.

In this article, we adopt an orchestrated spectrum sharing approach that integrates the distributed located UEs/DTV receivers, base stations (BS), DTV stations, and the Software-Defined Networking (SDN) controller into an amalgamated network with the real-time exchange through Internet. SDN is a new programmable network structure in light of the recent advancement of computing and network virtualization technology. It will be able to integrate the control entities of underlying infrastructures of broadcasting and wireless systems to a controller in control layer to realize global spectrum management. This way, the rigid network architecture that relies on vendor-specific configuration interfaces can be innovated into a programmable common interface for spectrum information exchange and configuration. Instead of indirect control of the heterogeneous networks, the controller is now aware of the real-time spectrum utilization and the unused spectrum resources are reported and pooled for sharing.

In order to effectively protect incumbent users and efficiently share the pooled spectrum resources, real-time spectrum information in terms of the 3D interference map will be considered for new spectrum access with the help of SDN’ global view. With distributed inputs reporting from incumbent systems, a 3D interference map is generated to guide the spectrum sharing procedure. New arrivals then access the shared spectrum only when the interference level is within threshold in order to guarantee overall network performance. Protected spectrum sharing improves spectrum utilization by allowing multiple users, both government, and commercial, to dynamically access the spectrum resource in a controlled way.

ABSTRACT. Radio Environment Map is a tool frequently envisaged to be one of the technical enablers for practical deployment of secondary systems operating in the legacy bands and following some of the spectrum sharing policies. However, the creation of such a map requires the detailed knowledge about the parameters of the incumbent system that has to be protected. The ways, how the information is obtained and stored in the system has to be accurate and cannot require involvement of too much resources, e.g., for collecting the measurement samples. Thus, in this paper we provide a discussion on the ways how the ordinary kriging algorithm can be applied for creation of an indoor and outdoor coverage map of the TV signal - a map that could be used by REMs for management of secondary transmission. Based on the real signal measurements conducted in Poznan, Poland, the map was calculated using various approaches, which are compared and discussed in the context of practical deployment of REM based systems operating in the TV-band, leading to better use of spectrum and converngence of broadband wireless and broadcasting systems.

ABSTRACT. In this paper we present a novel algorithm for moving object detection and segmentation, operating on H.264 bit streams. Compared to more traditional pixel-based approaches, the novelty of the algorithm consists of directly using the motion features embedded into the H.264 bit stream, thereby achieving real time operational capability. This makes the algorithm ready to be installed in any video surveillance system, enabling for better resource allocation and facilitating the deployment of distributed systems. The method we propose measures the statistical disorder of the motion vectors field at the boundary of the moving objects to achieve at the same time detection and segmentation. Further refinement of segmentation is achieved using temporal correlation of motion vectors. The algorithm has been tested on the traditional videos used to benchmark video compression algorithms, as well as on the iLids dataset.

ABSTRACT. There have been numerous suggestions that ATSC 3.0 be used to provide service targeted to mobile receivers as part of a broadcasters overall strategy. A physical layer pipe (PLP) optimized for mobile receivers will be readily receivable by rooftop antennas over a much larger area than today's 15 dB coverage contour [1], [2]. This will frequently result in new/additional coverage overlap between same-network TV station affiliates, with implications for program choice, competition, broadcaster- affiliate contracting, must-carry and retransmission consent fees. In this paper, we conduct a coverage propagation study to quantify the extent of this new competitive overlap.

In the U.S., there is not a single national TV media market; broadcast TV is distributed across a complex network of 210 Designated Market Areas (DMA), where most local TV stations are affiliated to a national broadcast network–e.g. ABC, CBS, NBC and FOX [3]. Between two local TV affiliates of the same network, most of the popular programming comes from their parent network, and is thus the same. Because they usually serve different local metropolitan areas, however, they differ in the provision of tailored local news and other forms of local programming. Due to station ownership limitations, neighboring local TV affiliates typically don't share the same owners, and therefore they may compete where viewership overlaps.

By their nature, DMAs, just like TV stations, are not perfectly isolated from each other. In the U.S., as opposed to other countries with more centrally planned TV spectrum policies, each TV broadcaster is free to decide–subject to frequency availability, maximum allowable transmit power, interference constraints and height limits–where to locate its antenna, which population to serve, and even the size of their coverage area. Thus, a significant level of inter-market competition exists due to overlaps between coverage areas, particularly between broadcasters that carry content from the most popular national broadcast networks such as ABC, CBS, NBC and FOX [3]. Increasing stations' coverage can provide viewers with more programming options due to increased overlap of over-the-air (OTA) coverage, but can also create frictions in existing local TV markets. In this analysis, we quantify the extent of these new frictions.

Unlike ATSC 1.0, the ATSC 3.0 standard will provide TV broadcasters with the ability to send simultaneous programs at a variety of signal robustness levels via different modulation and coding schemes (MCS). Thus, broadcasters can, from the same broadcast tower, serve both fixed and mobile receivers: e.g. traditional household TV sets, handheld devices and mobile phones. In fact, one of the top priorities for next-generation broadcast TV systems has been deploying high-quality mobile TV. For broadcasters, one barrier has been the additional 20 to 30 dB of link budget required by mobile receivers as compared to fixed rooftop receivers [1], [2]. To address this challenge, ATSC 3.0 has included Layered-Division-Multiplexing (LDM) as part of its physical layer (PHY) specification. It has been shown that LDM offers a much better channel capacity allocation than F/TDM when both fixed and mobile services with dissimilar bitrates and robustness levels are transmitted over a single RF channel [1], [2]. By properly configuring each services power levels, LDM allows single-transmitter broadcasters to provide a mobile coverage area that is closer to the traditional coverage area for fixed receivers, reducing the usual mismatch.

With ATSC 3.0, perhaps the most cost-effective scenario for deploying mobile services, [4], [5], is to use LDM in conjunction with Scalable Hierarchical Video Coding (SHVC). The primary program is sent over two layers for both fixed and mobile receivers. First, a SHVC base layer (BL) transmitted over the LDM upper layer (UL) is optimized for mobile reception, and second, a SHVC enhancement layer (EL) over the LDM lower layer (LL) is optimized for UHD resolution (or other enhancements such as HFR, Wide Color Gamut, etc.) at fixed rooftop receivers. Recent examples in the LDM literature show that with a 6 MHz channel, a TV broadcaster using ATSC 3.0 can provide enhanced quality fixed rooftop coverage at the same 15 dB SNR operating point that U.S. TV stations operate today, and at the same time, provide a mobile service coverage very close to that 15 dB coverage contour [1], [2]. At the same time, because of its more robust coding, the LDM UL alone can be received by fixed rooftop receivers at distances far beyond todays 15 dB coverage contour for TV reception [6]. As the BL is typically encoded at still relatively good resolution, e.g. at least 720p, using this mobile strategy can significantly increase a TV stations reach and viewership in a seemingly unintended way. As noted above, this can have important implications for program choice, competition, must-carry and retransmission consent negotiations.

In a first step, we will examine a number of representative ATSC 3.0 deployment scenarios that may project a stations coverage beyond today’s 15 dB contour. Specifically, we will assume either a high resolution UHD/4k video or multiple HD streams intended for rooftop receiver service, and a downscaled version —qHD or 720p/1080i HD— for mobile receivers. In order to preserve the area and population covered of existing TV stations, we will assume that all LDM LL coverage needs to match todays TV stations existing 15 dB ATSC 1.0 service areas. These representative scenarios will vary depending on a) the bitrate requirements for rooftop and mobile services, b) the assumed link budgets and MCS, and c) the bandwidth/power allocation between LDM layers. Secondly, we will assume that for each of these scenarios, all TV stations in the U.S. will adopt them, so for each case we will calculate the new coverage overlap both in terms of area and in terms of population served. In particular, we will pay close attention to those TV stations affiliated with the largest U.S. national broadcast networks. To do this, we use the Longley-Rice propagation model, [6] along with data publicly available in the FCC’s Consolidated Database System (CDBS) Public Searches [7], which centralizes information about location, authorized power, antenna height and horizontal transmit antenna pattern of all TV stations across the U.S. Census data is used for population distribution. With this data, we are able to calculate for all TV stations their area and population covered, and particularly the extent of newly created affiliate overlap.

ATSC 3.0 is substantially more flexible than ATSC 1.0. Broadcasters will use this flexibility to support a variety of business cases. Studies such as ours are essential in analyzing the inter-broadcaster competitive implications of alternative scenarios. Further, changes in population served resulting from various strategies may provoke regulatory response, as they can affect public values such as localism, must carry, and the contentious issue of retransmission consent.

REFERENCES [1] L. Zhang, Y. Wu, W. Li, H. M. Kim, S. I. Park, P. Angueira, J. Montalban, and M. Velez, “Channel capacity distribution of Layer-Division-Multiplexing system for next generation digital broadcasting transmission,” in 2014 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, pp. 1–6, June 2014. [2] L. Zhang, W. Li, Y. Wu, X. Wang, S. I. Park, H. M. Kim, J. Y. Lee, P. Angueira, and J. Montalban, “Layered-Division-Multiplexing: Theory and Practice,” IEEE Transactions on Broadcasting, vol. 62, pp. 216–232, March 2016. [3] Federal Communications Commission Report. In the Matter of Annual Assessment of the Status of Competition in the Market for the Delivery of Video Programming: Seventeenth Report, MB Docket No. 15-158, May 6, 2016 [4] C. Regueiro, J. Barrueco, J. Montalbn, U. Gil, I. Angulo, I. Eizmendi, P. Angueira, and M. Velez, “SHVC and LDM techniques for HD/UHD TV indoor reception,” in 2015 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, pp. 1–6, June 2015. [5] R. Chernock, “ATSC 3.0 Overview [Keynote],” in 2015 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting, June 2015. age and Interference Feb. 06, 2004. [7] FCC Media Bureau Filing Systems and Databases, Consolidated Database System (CDBS) Public Searches in FCC website, www.fcc.gov, 2013.

---------------- IEEE BMSB 2017

Corresponding Author:

Marvin A. Sirbu Professor, Carnegie Mellon University Baker Hall 129K 5000 Forbes Avenue Pittsburgh, PA 15213-3891, USA

Phone: +1 412 268 3436 FAX: +1 412 268 3757 email: sirbu (at) cmu.edu

ABSTRACT. In this paper, we investigate the problem of optimal content cache strategy with content of multi-available streaming rates over wireless network of different user channel condition. Our design objective is to maximize the quality of experience (QoE) at the user end under a limited cache server storage when considering the different channel condition among users. More specifically, the proposed strategy chooses the optimal subset of streaming rates based on channel condition and storage budget so that the expected QoE value at the user end is maximized. As a result, the propose strategy promote the utilization efficiency of insufficiency network resources and cache budget. Furthermore, through the analysis of the simulation results, there is a tradeoff between the cache size and the QoE value, depending on the number of users and their channel condition. The proposed cache strtategy is based on the MMT standard.

ABSTRACT. With the increased demand for multimedia content in the last decade, the television industry took a step forward and introduced the Hybrid Broadcast Broadband Television (HbbTV) standard. This step led to a seamless merge between DVB-related and IP-based services into a single terminal (e.g. TV, set top box), increasing the array of services which broadcasters can provide to the public and, at the same time, leading to a highly increased user experience. HbbTV is an open architecture which uses elements from already defined and mature standards such as OIPF, CEA, DVB, MPEG-DASH and W3C, making it easier to develop and deploy new services. Another advantage of this architecture is that it has all the needed hardware components already available so the process of developing new services is performed mostly on the software side. Video-on-Demand and live streaming are two of the most important services currently delivered through HbbTV, both very demanding in the aspects of quality, making them heavy bandwidth consumers. The main purpose of this work is to present a bandwidth usage management system for HbbTV services based on the user’s behavior. The presented approach is an on-top, fully automated solution that tracks and monitors user location/position and, based on the retrieved information, decides whether the bandwidth usage of an active HbbTV streaming service can be diminished by reducing the stream quality or even pause the service with no side-effects on the user’s quality of experience.

ABSTRACT. This paper introduces a utility-based network selection algorithm enabled, which takes into consideration hybrid unicast-multicast transmissions and balances energy consumption and quality for video deliveries in Dense Heterogeneous Networks. Besides taking into account the trade-off between throughput and estimated energy consumption of the mobile device, the proposed solution also considers a efficient usage of radio resources.

ABSTRACT. This paper is proposing some models for optimizing the Device-to-Device communication in the context of smart cities, by using spectrum sensing. Different levels of decentralisation – in computational intelligence, signal processing, power management and a greater percentage of open-source software – are discussed in a business-oriented perspective. A network of low power, inexpensive SDR devices is presented as solutions for implementing the communication partner discovery and spectrum selection as part of D2D communication in radio-congested urban environments. These practical solutions cover not only cost-effective SDR receivers but also local intelligence, with a spectacular add-on: mobility.

ABSTRACT. Overview:

Measuring video quality ensures that video operators can monitor and control the pipe in order to deliver consumers the desired quality of experience (QoE), and enable CODEC engineers to optimize the performance of their encoding algorithms. In the recent years, researcher Staelens et al [STAELENS], stated that while “subjective quality assessment” measured with mean opinion score (MOS) “is the most accurate method for obtaining human judgments on video quality,” they lament that subjective evaluations “are time-consuming, expensive, and require specialized expertise.” Also, “they cannot produce real-time quality ratings throughout a distribution network.” This echoes the findings of several researchers evaluating objective QoE metrics and the industry has been in search of an objective video quality metric that is easy to compute and that matches the expensive ‘golden-eyes’ subjective tests.

This paper briefly surveys metrics (peak signal to noise ratio and structural similarity variants) and then presents results of the new Video Multi-Method Assessment Fusion (VMAF) metric proposed [VMAF] by Netflix. The author and colleagues used the VMAF metric to measure the performance of the latest generation RealMedia video CODEC, RMHD, comparing it with results from subjective evaluation. The paper continues to discuss forward-leaning QoE metrics and approaches that can be used to measure how well consumer perception of content matches the transmitted reality, and how operators can measure audience response to content.

Part 1: Survey of objective video quality metrics

There are currently multiple ways to measure video quality:

• PSNR (Peak Signal-to-Noise Ratio): PSNR is a traditional signal quality metric, measured in decibels. It is directly derived from mean square error (MSE) or its square root (RMSE). The formula used is: 20 * log10 ( MAX / RMSE ) or, equivalently: 10 * log10 ( MAX^2 / MSE )

where the error is computed over all the pixels in the video with respect to a reference video.

PSNR successfully provided a numerical value when comparing an original input file with a coded output file, but didn’t match the ‘golden eyes’ tests when humans were put in a room. Variants of PSNR include: • Frame-averaged PSNR includes a temporal component to the content comparison • PSNR-HVS-M3 applies the comparison in the frequency domain. Good correlation with MOS results are claimed 4 by researcher Daede et al with appropriate selection of weighting factors to the DCT coefficients.

• SSIM (Structural Similarity Image Metric): SSIM is a still image quality metric introduced in 2004 by researcher Wang et al [SSIM]. It computes a score for each individual pixel using a window of neighboring pixels. These scores can then be averaged to produce a global score for the entire image with respect to a reference image. The original paper produces scores ranging between 0 and 1; however, this is commonly expressed in a non-linear decibel scale: -10 * log10 (1 - SSIM)

Algorithmic variant, MS-SSIM, handles multi-scale windows while FMS-SSIM is a fast implementation with constrained scales. This metric has been adapted for streamed video and commercialized as SSIMWave [SSIMWave].

• VQM (Video Quality Monitor): VQM is a commercial tool that requires no reference but measures the video quality with respect to the format and protocol specification. It can objectively measure blockiness, blurriness and frame rate.

• VMAF (Video Multi-Method Assessment Fusion): VMAF was specifically formulated by Netflix to correlate strongly with subjective MOS scores. It is a full-reference perceptual video quality metric that aims to approximate human perception of video quality [VMAF]. This metric is focused on quality degradation due to compression and rescaling. VMAF estimates the perceived quality score by computing scores from multiple quality assessment algorithms and fusing them using a support vector machine (SVM). Currently, three image fidelity metrics and one temporal signal have been chosen as features to the SVM – Anti-noise SNR (AN-SNR), Detail Loss Measure (DLM), Visual Information Fidelity (VIF) – and the mean co-located pixel difference of a frame with respect to the previous frame. The temporal component is important and is often lacking in metrics that merely compare a decoded image with a reference image. The rate control mechanism of a CODEC is continuously making a trade-off between spacial and temporal quality. VMAF is a good metric by which that compromise can be judged.

Part 2: Results VMAF used to measure CODEC performance

Chart of VMAF vs. MOS scores for RealMedia CODEC, RMHD. (This will be included in the final report)

VMAF is a definite step forward in producing an objective quality metric that correlates well with subjective testing. VMAF takes a fraction of the processing time of a video encode because it is specifically developed using machine-learning that was trained with ground truth derived from subjective testing. It has become an important metric in optimizing CODEC parameters and tuning rate control strategies that correlate well with ‘golden eyes’ testing.

Part 3: Broadening the scope of QoE

All aforementioned methods measure performance of the pipe from reference content to the consumer display. If we were to broaden the scope of the pipe, we also broaden our view of quality of experience. Let’s consider the start of the pipe as the reality captured by the camera. The end of the pipe is not the consumer display; rather, it’s perception of the content in the visual cortex. Namely, it’s about understanding how reality, represented by a sequence of the images, is perceived, and creating tools and metrics to prove how perception matches reality. Such a QoE regime will require a sophisticated model of human visual perception as well as 3D model of the original scene captured. We would want to measure audience response and then continuously train and re-train our QoE machine to adaptively provision the content delivery pipe to optimal consumer satisfaction.

So, what will make this possible? Current consumer technology delivers video content on mobile devices, which have cameras built into them. Simply put, a consumer watches a screen, and the screen is watching the consumer. Screens have now evolved to a point at which they are being used to take audience measurements—whether it’s counting the number of audience members, or estimating age and gender. What’s more, screens are now measuring an audience members’ facial expressions and doing so accurately—in multiple dimensions.

This is beneficial because consumers don’t need to click in or write comments to share feedback. For example, imagine a million people watching the presidential debate and companies measuring their feedback based on facial expression (e.g., are they frowning or smiling). This becomes an anonymized “heat map” of audience reactions on the content timeline.

Part 3: Real-time feedback loop

A real-time client-side measurement of audience satisfaction and response to content must be simple to execute and lightweight to transmit back as an analytics stream.

Facebook Live is one example—emojis fly by, representing reactions that previous people had when watching the same content. As such, this becomes very relevant to business and advertising companies.

Part 4: How that drives advertising systems

Facial expressions captured while viewers watch ad supported content would provide valuable feedback to both program content and commercials. Ads could be targeted in real-time to the audience in front of the screen with analytics automatically returned. Privacy concerns are not to be ignored. At CES 2017, Amazon’s Alexa won “Best Innovation,” a system that listens to your every word and transmits some data to the cloud. The public is becoming more accepting of such measurement in exchange for increased utility.

Conclusion

QoE metrics will evolve beyond measurements of bits per pixel, buffer overrun rate, and start time. The broader remit of service providers is to deliver reality to perception, bridging from a real scene to the viewer’s visual cortex. We consider broader technologies and techniques to measure and optimize the consumer satisfaction of an entertainment service.

###

ABSTRACT. Performance of a simplified transmitter/receiver structure based multi-carrier DS-CDMA with bi-orthogonal keying employing Adaptive Trellis Coded Modulation (ATCM) is analyzed. An efficient ATCM scheme using Generalized Symbol-rate-increased (GSRI)-ATCM has been proposed. However, the system structure requires multiple numbers of GSRI-ATCM encoders and decoders. This implies that the system complexity increases along with the increase of the number of sub-carriers. In this paper, a simplified structure by reducing the number of encoders and decoders are considered. The optimization of the method of selecting the coding rate and modulation index is proposed. We employ discriminant analysis method for selecting the coding rate and modulation index. The BER performance and throughput performance are evaluated and compared with those of the conventional scheme and the full adaptive scheme.

ABSTRACT. I. INTRODUCTION The Smart Grid is a recent paradigm applied to power systems, which proposes the support of the power networks with information and communication technology (ICT) systems, that can bring many benefits to the electrical systems in terms of global efficiency and quality of service of the distribution network. The liberalization of the energy market and the commitments under the Kyoto Protocol resulted into a strong drive towards a more sustainable production and utilization of the energy [1]. The engagement of governments in the direction of exploiting renewable energy sources (RES) for energy production, along with a resolute incentive policy, have fostered the spreading of a large number of distributed generation (DG) systems, frequently of various types and sizes, located on both public and private buildings, in rural and industrial areas. Transitioning from today’s traditional power grid to the smart grid of the future is the key to efficient wide-scale integration of distributed energy resources (DERs) into mainstream energy generation. Efficient, wide-scale, and cost-effective DER deployment requires that the system operator knows what devices are installed and where [2]. Along the main benefits that can be achieved by encouraging the development of smart grid [3]: - improving the management of load and generation variability, and reduce congestion and network inefficiencies, with a proper distribution/energy management systems (DMS/EMS); - reducing the number and duration of interruptions, especially by making it possible to operate in intentional islanding or with network reconfiguration, with smart control of the power line breakers. The first point can be classified as slow dynamics operation, since the time cannot be considered as the most important concern in the intervention performance. On the other side, the operation on the breakers is highly time critical, since it is necessary to guarantee an instantaneous intervention on the breakers in order to assure an efficient intervention, for example during or after a fault extinction. The implementation of such systems requires a robust infrastructure that allows interrupting the power flow when faults occur, to define the area of a fault, while minimizing the disconnection of loads and avoiding that the fault may be fed by RESs. For these reasons, the ICT in this context is not only an upgrade to the system equipment, but a fundamental requirement for supporting the distribution network operation, monitoring and control. Recently, a combination of several cellular technology advances have emerged that will most likely bring the debate to a close and position LTE-A-based Fifth-Generation (5G) mobile wireless technology that will be supported by emerging internet of things (IoT) as the global future grid communication networking standard [2]. Communication systems are crucial for the correct implementation of smart Grid in power distribution systems and they have to be designed taking into due account the requirements of expected functionalities such as network operation with cooperative DER, system protection and/or post fault network reconfiguration, taking strictly into account quality, low latency, reliability, traffic safety, and noise immunity [4]. 5G radio access technology (RAT) meets the requirements of smart Grid applications. Moreover, 5G is intrinsically built for diverse new applications based on machine-to-machine (M2M) communication and for control and monitoring of critical infrastructures like smart Grid [5]

II. RELATED WORKS Various works in literature propose different approaches for the smart Grid monitoring and control. In [6] an extensive overview of the state of the art on distribution system control architectures is presented. Conceptually, the frameworks are classified into three main categories: centralized architecture, hierarchical architectures and decentralized architectures. The advantages of hierarchical and decentralized over centralized approaches are highlighted. In particular, it is observed the tremendous increase in both dimension and complexity of the distribution system so that an effective management structure should benefit from the use of a more flexible control and management architecture. In [7] the authors present a comparison between a number of wireless sensor networks (WSN) architectures for energy consumption monitoring in the smart grid. The aim of this paper is to highlight the potential benefits of having distributed storage and distributed information processing, in comparison to centralized approaches, in terms of network performance. The performance of the network is evaluated considering energy consumption and communication cost, processing requirement, storage requirement, bandwidth requirement, traffic and financial cost of the different architectures. In [8], Wei et al. proposed a study on delay performance in distributed system versus centralized system, and the impact of asynchronous message delivery based on a micro smart grid. The experimental results show that for computationally intensive applications, distributed control system causes a worse delay performance compared to centralized control system. In particular, the centralized control system outperforms the distributed control system by 50% in their case study.

III. PROPOSED APPROACH Both slow dynamics and fast dynamics operation can be performed with two different logics: Centralized and Distributed Network Management (Figure 1). The centralized network management consists in a central DMS/EMS that checks the data that are sent from the measurement unit (MU), and performs an operational intervention over the Intelligent Electronic Devices (IEDs) that are installed on generators, breakers, active loads, storage units, in order to maintain the network in optimal regime. The information data flow is vertical, that means that each MU and IED exchange data only with the central controller. On the other hand, on a Distributed Network Management scheme each IED and MU are associated with an agent, which is a local control algorithm that can communicate with the neighbor devices in order to assess the network status and perform the operational actions to keep the optimal conditions and solve any network contingency. The information data flow is horizontal. A local data aggregator may act as interface between the local agents and the Distribution System Operator (DSO) in order to coordinate the network operation. This work will focus on fast dynamics operation, since it is one of the most interesting aspects to evaluate the suitability of a novel technology for smart Grid support. An innovative automatic reconfiguration algorithm that exploits the 5G features represents the main novelty proposed, allowing to improve the global reliability by increasing the resilience of the power network. Indeed, the 5G cellular network enables utilities to wirelessly connect to the entire distribution grid assets. Associated with each of these assets, there is one or more low cost stationary LTE-A-enabled M2M module (embedded chip set) that sends and receives data to and from the macro base stations (BSs), small cells, or another M2M module [2]. Within technologies for M2M, the LTE-A [9] and the NarrowBand IoT (NB-IoT) guarantee several advantages compared to other communication standards (i.e., ZigBee, BLE, LP-WiFi, 3GPP Rel. 8) such as low latency (i.e., up to 10 ms), service level agreement, scalability, reliability, low power, and large coverage. Nowadays, a standard for distribution network control and monitoring is fully or partially missing. According to the advantages above mentioned, the 5G architecture based on LTE-A and M2M can be considered a strong candidate for data exchange between smart grid devices for monitoring and control. In our work we consider a M2M chip set embedded on each IED in order to ensure a horizontal bidirectional data flow. This framework has been applied on a case study where network contingencies are solved in a distributed control architecture. The proposed solution is validated with simulations performed with a software tool that consists of DigSilent PowerFactory and OMNeT++, integrated in a co-simulation framework. Different smart grid scenarios will be presented, in order to analyze the global performances of the smart grid from both electrical and communication point of view.

REFERENCES [1] L. Bird, M. Bolinger, T. Gagliano, R. Wiser, M. Brown, and B. Parsons, “Policies and market factors driving wind power development in the United States,” Energy Policy, vol. 33, no. 11, pp. 1397–1407, Jul. 2005. [2] M. A. Ali and A. Mohamed, “5G Cellular Technologies for Supporting Future Power Grid Communication Networks - IEEE Smart Grid,” IEEE Smart Grid Newsletter Compendium, Oct-2016. [3] N. S. Wade, P. C. Taylor, P. D. Lang, and P. R. Jones, “Evaluating the benefits of an electrical energy storage system in a future smart grid,” Energy Policy, vol. 38, no. 11, pp. 7180–7188, Nov. 2010. [4] M. Garau, G. Celli, E. Ghiani, F. Pilo, and S. Corti, “Evaluation of Smart Grid Communication Technologies with a Co-Simulation Platform,” IEEE Wireless Communications Magazine. Special Issue on Smart Grids., Apr-2017. [5] D. Moongilan, “5G wireless communications (60 GHz band) for smart grid #x2014; An EMC perspective,” in 2016 IEEE International Symposium on Electromagnetic Compatibility (EMC), 2016, pp. 689–694. [6] M. A. Matos et al., “Control and Management Architectures,” in Smart Grid Handbook, John Wiley & Sons, Ltd, 2016. [7] R. M. Oviedo, F. Ramos, S. Gormus, P. Kulkarni, and M. Sooriyabandara, “A Comparison of Centralized and Distributed Monitoring Architectures in the Smart Grid,” IEEE Syst. J., vol. 7, no. 4, pp. 832–844, Dec. 2013. [8] M. Wei and W. Wang, “Toward distributed intelligent: A case study of peer to peer communication in smart grid,” in 2013 IEEE Global Communications Conference (GLOBECOM), 2013, pp. 2210–2216. [9] “3GPP Release 13 - http://www.3gpp.org/release-13.” [Online]. Available: http://www.3gpp.org/release-13.

ABSTRACT. I. INTRODUCTION Powerful mobile computing devices such as smartphones and tablets are ubiquitous, and their rapid adoption by large portions of the population have led to massive changes to the way in which the WWW is accessed and the types of traffic that traverse networks. Video content has now become the dominant traffic type on the Internet and is forecast to further increase is market share [1]. Most smartphones are now multi-homed devices having both cellular and WiFi capable interfaces enabling the user to connect to the Internet over a variety of wireless networks.

The amount of available bandwidth within wireless networks fluctuates greatly depending on user location, physical environment, number of current users, etc. and this variation in available bandwidth makes video streaming difficult. Adaptive video content streaming over HTTP was developed to deal with the problem of variations in the amount of available bandwidth. HTTP video on demand streaming is a combination of downloading and concurrent playing of the video. When a sufficient amount of content has been downloaded, requiring some delay, the video is played from the buffer. Dynamic Adaptive Streaming over HTTP (DASH) [2, 3, 4] is widely used, the basic idea is to break the video content into segments of between 2 to 10 seconds duration and to encode each of these segments at multiple bit-rates or quality levels. The DASH player (client) then requests the video content at the quality level that best matches the current available bandwidth at the user’s location using the standard GET command. This approach enables the DASH player to maximise the use of available bandwidth and to alleviate the effects of changing network conditions on the video stream.

II. RELATED WORKS

The Quality of Service (QoS) provided by a communications network can be objectively described by certain measurable network performance parameters such as jitter, delay, packet loss, etc. Satisfactory QoS does not ensure that all service users will perceive the service as being ‘good’. Quality of Experience (QoE), perceived quality, on the other hand, is subjective and considers the end-users perception of the overall value of the service. In the case of HTTP streaming [5, 6] show that initial delay and video stalling are primary factors influencing QoE.

In any multimedia streaming system some initial delay is unavoidable since a portion of the video content must be downloaded prior to decoding and the start of playback. Although it is typical for the initial delay to be somewhat longer than is strictly necessary from a technical point of view initial delay has a small impact on perceived quality related to the length of the delay. The authors of [7] demonstrate that initial delays of up to 16 seconds will be tolerated by users and that up to 90% of users prefer some initial delay to stalling.

Stalling, the pausing of video playback due to the playback buffer being depleted faster than it can be filled, has a significant negative impact on perceived quality [8]. It is considered to be worse than frame rate reduction [9], multiple short stalling events are considered to be worse than a single stalling event [8], stalling at irregular intervals is worse than stalling at regular intervals [9], etc. All in all stalling should be avoided where possible since it results in highly dissatisfied end-users. Even a small amount of stalling severely degrades perceived video quality.

A great deal of work has taken place in the area of adaptive streaming over HTTP. Many different approaches have been taken but they all depend on similar information such as available throughput during downloading or buffer levels as inputs into their decision making processes. The decision under consideration is whether or not to select a higher or lower resolution DASH segment for the next download. If the buffer level is critically low then a lower bitrate segment is selected for down load since the throughput over the wireless link cannot keep pace with the buffer playout rate. Selecting too high a bit-rate will result in stalling since the buffer will become depleted before it can be replenished. Selecting too low a bit-rate will result in a very noticeable degrading in the frame resolution.

III. PROPOSED SOLUTION

Various strategies have been developed to try and prevent stalling of video streams, these strategies focus on changing the selected segment bitrate in response to reductions in throughput. We propose a novel system that leverages the multi-homed nature of modern mobile devices themselves, which instead of avoiding stalling events in the video playback by downgrading the segment bitrate implements a strategy of activating a second connection channel be used to download the necessary segments and increase throughput temporarily [Figure 1].

Figure 1 Single vs Dual Channel Segment Requests

Previous work has demonstrated that under certain circumstances it is advantageous for a mobile user to switch connectivity from a cellular network to a WiFi network. In this situation the cellular interfaces data connection is typically shutdown to conserve energy and to preserve the end users data cap. Our proposed system monitors the playout buffer levels and the achieved throughput on the WiFi downlink, if the buffer cannot be replenished quickly enough using a single connection to avoid the video stalling the dormant cellular data connection is re-established to increase overall throughput to the buffer [Figure 2]. The throughput achieved on the primary (WiFi) interface is monitored and when conditions improve to the point at which it can sustain replenishing the playout buffer on its own the cellular data connection is made dormant again.

Figure 2 Proposed Solution Architecture

This approach has the benefit of not requiring a reduction in segment bitrate while at the same time minimizing the total amount of data downloaded over the cellular interface thus reducing the load on the cellular network and protecting the end users data cap. It also enables the user to greatly reduce the amount of delay caused by the initial filling of the playout buffer. Reducing initial delay, maintaining a higher bitrate and reducing the risk of stalling improves QoE for the end user.

IV. TESTING PLAN

The test environment will consist of a Lenovo ThinkPad laptop equipped with an Intel i7 processor, 16GB of RAM and an SSD drive running Debian 8 GNU/Linux. The proposed solution will be tested using the NS3 [10] simulation package and Linux LXC containers [11]. Two Debian based LXC containers will be created as well as a set of tap/bridge devices to enable the external containers to send traffic to and from the NS3 network simulation. One of the Debian containers will run a modified version of VLC and acted as a streaming server while the other container hosted a VLC client. Simulations are to be conducted in two phases and the results compared. In Phase 1 the mobile host will stream a video formatted for MPEG DASH without modifying its behaviour using a single wireless interface. During Phase 2 the client will implement the proposed solution and employ one or both wireless interfaces as necessary depending on the buffer state and the current available throughput.

REFERENCES

[1]Cisco, “The Zettabyte Era: Trends and Analysis”, White Paper, July 2016, https://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/vni-hyperconnectivity-wp.html, Accessed: September 2016

[2]ISO, “Information technology -- Dynamic adaptive streaming over HTTP (DASH) -- Part 1: Media presentation description and segment formats”, ISO/IEC 23009-1:2012, April 2012

[3]A. Vetro, I.Sodagar, “The MPEG-DASH Standard for Multimedia Streaming Over the Internet”, Industry and Standards, IEEE Computer Society, October – December 2011 [4]T. Stockhammer, “Dynamic Adaptive Streaming Over HTTP: standards and design principles”, MMSys ’11 Proceedings of the Second Annual Conference on Multimedia Systems, Pages 133-144, 2011

[5]T.Hobfeld et al., “Quantification of YouTube QoE via crowdsourcing”, in Proc. IEEE Int. Symp. Multimedia, Dana Point, CA, USA, 2011, pp. 298-303.

[6]R.K.P. Mok, W.W. Chan, X. Luo, and R.K.C. Chan, “Inferring the QoE of HTTP video streaming from user-viewing activities”, in Proc. ACM SIGCOMM W-MUST, Toronto, ON, Canada, 2011, pp. 31-36

[7]T. Hobfeld, et al., “Initial delay vs. interruptions: Between the devil and the deep blue sea”, in Proc. 4th Int. Workshop QoMEX, Yarra Valley, Victoria, Australia, 2012, pp. 1-6

[8]Y. Qi, M. Dai, “The effect of frame freezing and frame skipping on video quality”, in Proc. 2nd Int. Conf. IIH-MSP, Pasadena, CA, USA, 2006, pp. 423-426

[9]Q. Huynh-Thu, M. Ghanbari, “Temporal aspect of perceived quality in mobile video broadcasting”, IEEE Trans. Broadcasting, vol. 54, no. 3, pp. 641-651, September 2008

[10] NS-3, www.nsnam.org, accessed Dec. 2016 [11] Linux Containers, https://linuxcontainers.org/, accessed Dec 2016

ABSTRACT. Fully Convolutional Network (FCN) makes it possible for deep learning to deal with image semantic segmentation. But the results are still coarse in pixel level. Therefore, probabilistic graphical models such as Conditional Random Fields (CRFs) have been developed as effective methods to obtain more contextual information. However, existed works still have poor performance under high noise disturbance and low-resolution conditions. Based on the existed fully connected pairwise CRF models, we propose a new type of potential, Object Clique Potential, which is integrated into the CRF model to further enhance its optimization performance. By including this potential in a CRF embedded within a CNN, significant improvement has been achieved over joint CNN and CRF systems which only use unary and pairwise potentials. Under experiment, this model can be further adapted to dealing with semantic segmentation in low-resolution video processing. The proposed model has been testified on the problem of semantic image segmentation of the challenging Pascal VOC 2012 segmentation benchmark and reaches the state-of-the-art level segmentation results.

ABSTRACT. In this paper we consider a novel method based on multihop Device-to-Device (D2D) communications where proximity links are established among User Equipments (UEs) in proximity in a Long Term Evolution-Advanced (LTE-A)-based network. The aim of the proposed solution is to improve the multimedia multicast sessions and transmissions in terms of throughput and mean downloading time per user by providing ubiquitous and reliable connectivity to the larger number.

ABSTRACT. In this paper we propose an optimal solution regarding the multicast subgroup configuration problem for providing multimedia services over 5G satellite systems. In particular, the so-called Optimal Multicast Subgroup Configuration (OMSC) approach aims at maximize the aggregate data rate with an execution time that is sensibly smaller compared to other solutions available in literature.

ABSTRACT. This paper presents the scenario of terrestrial broadcast services, the needed adaptations for the next generation television and proposes an adapted ISDB-T system which makes use of diversity at reception through the MRC method as well as uses LDM, LDPC and SDR/GRC, in order to improve performance, spectral efficiency and useful bit rate to UHDTV applications, while using a layer still compatible with the traditional ISDB-T system.

ABSTRACT. In this paper, we propose an efficient FFO(fractional frequency offset) estimation method for OFDM systems with RP(Roll-off period). RP is used to transmit more subcarriers in a limited bandwidth. Because neighboring OFDM symbols overlap in RP region, RP makes noises in CP(cyclic prefix) and degrades FFO estimation when Beek’s method is used. The proposed method estimates FFO based on the Beek’s method by detecting and eliminating RP from CP. We showed the test results of RP detection and explained the proposed FFO estimation.

ABSTRACT. As human is the final receptor of images, user-perceptive quality of images need to be scaled by subjective quality assessment method, which is also important for learning a high performance of objective image quality assessment model. There are characteristics of randomness and individual difference among subjective testers. Those scores made by unreliable subjects that may affect the scaling of image quality scores, are defined as outliers. Recent studies have shown that pairwise comparison is more reliable and accurate than single stimulus method. However, the traditional outlier detection method for pairwise comparison, i.e, transitivity satisfaction, only compared results for individual participant, ignoring the statistical information among subjects for each pair of images. In this paper, an improved scaling method for pairwise comparison was proposed, which combines transitivity satisfaction and statistical information among all participants for each pair of images to evaluate the reliability of participant and reject the outlier. Pairwise comparison experimental result on a geometric distorted stereoscopic image database demonstrates the improvement of the proposed method.

ABSTRACT. With the rapid development of media services, it’s an urgent demand that videos uploaded to the Internet need to be processed with high efficiency and high quality. We set up a lightweight distributed system for media processing to address this problem. The architecture of our system consists of three layers, which are connected by three flows. This system has been deployed for a public video processing service for UGC in a private cloud environment. Moreover, the functional extension and efficiency optimization are discussed in this paper. Overall, the deployed public service shows remarkable performance in both processing efficiency and network load.

ABSTRACT. The fifth generation of mobile communications (5G) regards Internet of Things (IoT) as one of the main force driving the evolution of mobile communications to achieve the aim of “internet of Everything” towards 2020. Due to channel variety and diverse service requirements, highly flexible rate -compatible channel codes with variable lengths are particularly required in 5G. As one of the big challenge, the continuity of the code rate and the bit-by-bit variation of the code length are the ultimate demand of the rate/length-compatible codes’ proposal. Low-density parity-check(LDPC) codes have attracted wide attentions because of it’s low decoding complexity and capacity-approaching performance as well as high throughput. These advantages can exactly meet the demands for multi-rate multi-length channel codes. This paper proposes a design of rate-compatible length-compatible LDPC codes which is based on the idea of progressive extension of both information node and check node. Raptor-like structure and QC-LDPC codes are adopted to generate such a family of codes with fine granularity for all produced code rate and code length. We also use edge switch algorithm to search sub-optimal edges based on DE analysis result which can guarantee capacity-approaching performance. In addition, the proposal is simple and easy for hardware implementation.

ABSTRACT. Recent years have been characterized by the new so-called 4K format, also known as 2160p, which is part of the Ultra-high Definition television (UHDTV) standard defined by ITU-R BT.2020 [1]. Nowadays, together with the increasing growth of devices with ever greater display resolution, video-on-demand services such as Vimeo [2], YouTube [3], Netflix [4] and Amazon Instant Video [5] all offer 4K video contents with impressive detail and clarity. The annual worldwide shipments of UHDTVs are expected to grow nearly 719% over the next several years according to IHS’ forecasts, from nearly 12 million in 2014 to nearly 96 million in 2019, with over 300 million in use by the end of 2019 [6]. The UHD Alliance vision consists on providing consumers a new differentiated entertainment experience that delivers a premium expression of creative intent using next generation audio-visual technologies. However, UHD Alliance does not provide methodologies for subjective quality assessments [7]. As is known, Quality of Experience (QoE) of the consumption of multimedia content has great attention, both from academia and from industry. In order to assess video quality or to develop models that can estimate video quality, subjective tests are necessary. These tests need to be performed adhering to certain methodology that ensures that their results are valid. According to display resolution and emerging Ultra-high Definition (UHD) contents, QoE requirements represent an important issue. However, until now, there is no ITU recommendation concerning subjective tests in UHD resolution. The proposed paper will: • Review ITU recommendations that define video parameters of SD, HD and UHD resolutions. • Review ITU recommendations that describe the methodology to assess video quality in SD and HD resolutions. • Review reports found in the literature of subjective tests performed for UHD. • Based on the above-mentioned points, and on our own experience, propose some points that should be included in a recommendation that would describe the methodology to perform subjective tests to assess video quality in UHD resolution.

II. RELATED ITU RECOMMENDATIONS Recommendations ITU-R BT.601 [8], ITU-R BT-709 [9] and ITU-R BT-2020 [1] describe the parameter values for SDTV, HDTV and UHDTV, respectively. The Recommendation ITU-R BT.1129 [10] provides subjective assessment of standard-definition digital television (SDTV) systems. In particular recommends that the general methods for the subjective assessment of digital systems be as described in Recommendation ITU-R BT.500 [11]. This last defines general laboratory and home viewing conditions. Viewing distance and screen size selected in order to satisfy the preferred viewing distance (PVD) are valid for both SDTV and HDTV. The Recommendation ITU-R BT.710 [12] deals with subjective assessment methods for image quality in high-definition television and should also be made following the general methodology given in the above-mentioned ITU-R BT.500. The Recommendation ITU-R BT.2245 [13] on HDTV and UHDTV test materials for assessments of picture quality, simply lists video clips to use in the subjective tests, and the format specifications in terms of resolution, image size, aspect and sampling ratio, bit depth, quantization, system colorimetry and file format, both for 4K and 8K. Moreover, this recommendation describes the camera characteristics used for the capture of the material, in terms of image sensor, effective pixels, color definition, shooting mode, ISO speed range and shutter speed. However, subjective assessment methods for UHDTV are not provided in any ITU-R BT recommendation. III. REPORTS ON SUBJECTIVE TESTS Some studies were conducted for subjective QoE video quality. In [14] the authors contribute with a new 4K video dataset, consisting on 10 sequences each of them compressed at 6 different bitrates in HEVC/H.265 standard. The database includes full subjective Mean Opinion Scores (MOS) obtained from subjective tests. Zhu et al., also proposed and validated a new parametric model based on content clustering analysis. This work proposes a bit rate estimation model based on video content as a benchmark that can predict the bit rate required for encoding processing at a certain video quality. In [15] Cheng et al., proposed a subjective QoE model based HEVC encoder adaptation scheme for multi-user video streaming. The authors used several sequences with different resolution, quantization parameter and frame rate. The impact of HEVC encoders on video streaming has been evaluated in terms of packet loss rate (PLR) and MOS perceived by end users. Performances analysis are related to a limited set of 4K sequences. In [16] the authors deal with subjective quality assessment comparing UHD and HD resolution in HEVC transmission chains. The first chain is characterized by UHD source and displaying a UHD and HD obtained with a down-sampling with factor 2. The second chain starts with a UHD and a down-sampling with factor 2 in order to obtain a HD source that will display in one case at the same HD and in the second case with up-sampling with factor 2 in order to obtain the native UHD display. The work in [17], presented extensive subjective quality assessment results for HEVC-encoded 4K-UHD video contents in terms of two viewing distances of 0.75H and 1.5H, two color formats of YUV444 and YUV420, and three target bitrates of 18, 23 and 36 Mbps, and shows a possible feasibility of using the HEVC codecs for future 4K-UHDTV broadcasting. Finally, in [18] Tan et al. present subjective and objective results of a verification test in which the performance of the new HEVC standard is compared with its highly successful predecessor, the Advanced Video Coding (AVC) video compression standard (ITU-T H.264 and ISO). The HEVC verification tests were carried out for four categories of spatial resolutions: UHD, 1080p, 720p, and 480p. In these above mentioned works, the laboratory is set up according to [11] where the viewing distance, screen size and PVD curve are valid both for SDTV and HDTV but are not standardized for UHDTV. Moreover, the maximum observation angle relative to the normal, monitor processing and resolution are referred to consumer Cathode Ray Tube (CRT) displays. In order to make comparable experiments, new standards are needed to make subjective tests with UHD, which shall define the appropriate environment and lab setups for this new video resolution. IV. IN SUMMARY ITU-R BT series defines general laboratory and home viewing conditions with particular emphasis in PVD valid for both SDTV and HDTV whereas subjective video quality (SVQ) assessments for UHDTV are actually fully or partially missing. Based on existing ITU recommendations, on the methodology employed to perform the above mentioned subjective tests by different laboratories, and on the authors own experience, in the full paper the authors will propose some points that should be included in a recommendation that would describe the methodology to perform subjective tests to assess video quality in UHD resolution. As shown in [19] the authors faced with subjective tests for both SDTV and HDTV. Until now, the subjective quality assessments for UHDTV videos were conducted using SDTV and HDTV guidelines. This fact motivates our work in the full paper that will deepen the description of the methodology employed in each case in terms of distance view, horizontal and vertical angle view relative to the normal taking into account LCD screen, user classification, and home or laboratory environment. REFERENCES [1] “ITU-R Recommendation BT.2020-2-201510-I - Parameter values for ultra-high definition television systems for production and international programme exchange.” . [2] “Vimeo: Watch, upload, and share HD and 4K videos with no ads.” [Online]. Available: https://vimeo.com. [3] “YouTube.” [Online]. Available: https://www.youtube.com/. [4] “Can I stream Netflix in Ultra HD?,” Help Center. [Online]. Available: https://help.netflix.com/en/node/13444. [5] “Amazon.com: Amazon Video: Amazon Video.” [Online]. Available: https://www.amazon.com/Amazon-Video/b/ref=topnav_storetab_atv/166-1865049-8673925?_encoding=UTF8&node=2858778011. [6] “One-Third of US Households Will Have 4K TVs in 2019, IHS Says | IHS Online Newsroom.” [Online]. Available: http://news.ihsmarkit.com/press-release/technology/one-third-us-households-will-have-4k-tvs-2019-ihs-says. [7] “UHD Alliance.” [Online]. Available: http://www.uhdalliance.org/. [8] “ITU-R Recommendation BT.601-7-201103-I - Studio encoding parameters of digital television for standard 4:3 and wide screen 16:9 aspect ratios.” . [9] “ITU-R Recommendation BT.709-6-201506-I - Parameter values for the HDTV standards for production and international programme exchange.” . [10] “ITU-R Recommendation BT.1129-2 - Subjective assessment of standard definition digital television (SDTV) systems.” . [11] “ITU-R Recommendation BT.500-13 - Methodology for the subjective assessment of the quality of television pictures.” . [12] “ITU-R Recommendation BT.710-4 - Subjective assessment methods for image quality in high-definition television.” . [13] “ITU-R Recommendation BT.2245-1-2014-PDF-E - HDTV and UHDTV test materials for assessment of picture quality.” . [14] Y. Zhu, L. Song, R. Xie, and W. Zhang, “SJTU 4K video subjective quality dataset for content adaptive bit rate estimation without encoding,” in 2016 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), 2016, pp. 1–4. [15] Z. Cheng, L. Ding, W. Huang, F. Yang, and L. Qian, “Subjective QoE based HEVC encoder adaptation scheme for multi-user video streaming,” in 2016 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), 2016, pp. 1–6. [16] K. Berger, Y. Koudota, M. Barkowsky, and P. L. Callet, “Subjective quality assessment comparing UHD and HD resolution in HEVC transmission chains,” in 2015 Seventh International Workshop on Quality of Multimedia Experience (QoMEX), 2015, pp. 1–6. [17] S. H. Bae, J. Kim, M. Kim, S. Cho, and J. S. Choi, “Assessments of Subjective Video Quality on HEVC-Encoded 4K-UHD Video for Beyond-HDTV Broadcasting Services,” IEEE Trans. Broadcast., vol. 59, no. 2, pp. 209–222, Jun. 2013. [18] T. K. Tan et al., “Video Quality Evaluation Methodology and Verification Testing of HEVC Compression Performance,” IEEE Trans. Circuits Syst. Video Technol., vol. 26, no. 1, pp. 76–90, Jan. 2016. [19] J. Joskowicz, R. Sotelo, M. Juayek, D. Durán, and J. P. Garella, “Automation of Subjective Video Quality Measurements,” in Proceedings of the Latin America Networking Conference on LANC 2014, New York, NY, USA, 2014, p. 7:1–7:5.

ABSTRACT. In recent years, the late-model convolutional neural networks and feasible online datasets provide saliency detection with rapid development. Under the circumstances, several aspects in computer vision can incorporate saliency models, such as semantic segmentation, object detection, object proposals, image clustering and retrieval, cognitive saliency applications, video compression, video summarization, and video segmentation where extracted saliency maps are applied to pre-processing step. However, less attention is paid to videos because convolutional neural networks seem to make insignificant improvements to extract inter-frame information. In this paper, a novel saliency detection architecture is proposed for saliency detection in dynamic scenes combining two stream convolutional neural networks and recurrent connections from global-view to local-view. We enforce motion information in optical flow maps and appearance information in video sequences as significant cues.

ABSTRACT. The knowledge about parameters of digital signal in real time is very important to know the behavior of signal in the broadcasting process of Digital Terrestrial Television (DTT). LACETEL, Research and Development Telecommunications Institute, have been involved in the assimilation process of technologies linked to adoption by Cuba of the DTMB (Digital Terrestrial / Television Multimedia Broadcasting) as transmission standard. In this process we have developed an Intellectual Property (IP) module to obtain objective information about signal's behavior. In this paper is presented a DTMB signal monitoring tool based on a commercial STB (Set-Top Box), which operates with an property software developed over an open source eCos (embedded configurable operating system) system, using the STB because is a low capacity computer (256 MBytes) and have real-time system. The developed software is optimized to work in DTMB standard mode 6, whose characteristics are: 3780 carriers, FEC (Forward Error Correction) 0.6, 64QAM modulation (Quadrature Amplitude Modulation), PN420 header and 18,274 Mbps throughput. The Measurement Tool development consist of the inclusion of an intellectual property module in the firmware of digital television receiver, with the aim to include new features and functionalities to be used as a monitoring tool for signal strength and signal-to-noise ratio of digital television signal with DTMB modulation with 6 MHz channel. The work is divided into three parts, the first one describes the functionalities of IP module and how the calibration and correction process of the obtained values was performed. It explains in detail the functionalities of IP module: Integration of equations to get the signal strength in dBm from information provided by the Set Top Box; it includes features to store the strength obtained from the received signal every 60 seconds, 60 minutes and 24 hours; Inclusion of a function to fix the obtained SNR values ; Storage of the SNR parameter every 60 minutes; it also includes functionalities to calculate the maximum, minimum and average parameters of signal strength and SNR in 60 minutes and 24 hours. The second part explains the process to show the results in two new OSD interfaces added to the Set Top Box, the information is showed in graphs and tables form, and includes an Event Log that can be used to know the signal´s behavior in a time interval without the need of a technical specialist´s tracking. The third part is about to export the whole obtained information as a text file to an USB removable device. The event log is able to store the date and time and a description. Four types of events have been declared: signal loss, signal return, 2dB level changes, and alert when the signal level drops below -80dBm. This part also includes a comparison between obtained values from the Set Top Box and a professional measurement equipment for signal strength and SNR parameters.

ABSTRACT. Doubly selective channel estimation for a dynamic sparse channel in the large-scale multiple-input multiple-output (MIMO) systems has become a challenging problem. Considering the two parts: the large number of channel coefficients for MIMO systems requiring unaffordable pilot overheads, and the special feature of dynamic sparse channels: path delays change over time, we propose a differential block simultaneous orthogonal matching pursuit (DBSOMP) algorithm based on joint multi-symbol channel estimation to estimate dynamic channel parameters accurately. Taking advantage of the complex basis expansion model (CE-BEM), we can transform the original goal into estimating a series of jointly sparse CE-BEM coefficient vectors. Considering the common sparsity of the coefficients in different BEM orders among different antennas, we firstly arrange the coefficients in the way of block structure. The channel information corresponding to continuous OFDM symbols has temporal time correlation. Then based on superimposed pilot design, we apply DBSOMP algorithm to estimate the dynamic sparsity of each time slot and apply the actual sparsity on channel estimation. Simulation results demonstrate that the proposed DBSOMP algorithm achieves better performance than the conventional BSOMP algorithm.

ABSTRACT. Moving segmentation comes first in the pipeline of video analysis system. Although pixel domain methods are fully exploited, compressed domain methods still remain few relatively, especially for HEVC. This paper proposes a novel moving segmentation method in HEVC compressed domain based on logistic regression which can handle with camera jitter situation. Firstly, distinction between HEVC syntax elements in moving foreground and background is fully analyzed. F-score evaluation is performed to confirm the discrimination capacity of these features. Then these raw features, including prediction mode, MV, etc., are pre-processed into more useful and robust features. Noise is reduced by spatio-temporal filter and global camera motion is removed in this step. Afterwards, moving segmentation is modeled as a binary classification problem which takes the pre-processed features as input. To solve the classification problem, logistic regression is utilized in this paper. Experimental results present that the proposed method achieves comparable performance with state-of-art pixel domain method and 16.78 times faster processing speed. Nowadays, numerous video data is generated every minute, especially from surveillance systems. Although most video content is compressed with encoding standard, such as HEVC [1], analyzing these huge amounts of video data is still a great challenge. To solve this challenge, moving segmentation plays an important role for its first place in the queue of video analysis. The general idea of moving segmentation is to automatically generate a binary mask which divides the set of pixels into the set of moving foreground and the set of background pixels. Based on the segmentation result, subsequent video analysis such as moving object detection and tracking can be performed faster and better. Plenty of conventional moving segmentation algorithms are proposed, mostly executed in pixel domain [2-4]. For these methods, original video frames must be reconstructed by performing decoding, which is time- and memory- consuming. On the contrary, bit stream syntax elements such as motion vectors (MVs), DCT coefficients and block coding modes can be obtained with minimal effort. Thus, a number of compressed domain moving segmentation algorithms have been reported recently. Major of the works are based on H.264/AVC video coding standard [5-8]. HEVC is the latest video coding standard, providing higher compression performance compared to the previous standard H.264/AVC by 50% of bitrate saving at similar perceptual image quality. It is the trend that increasing amount of video will be encoded with HEVC. However, compared with H.264/AVC, there are only few literatures on moving segmentation in HEVC compressed bit stream [9-10]. [9] makes use of MVs, coding unit and prediction unit of HEVC to calculate the feature: Motion Intensity Count (MIC). With MIC, moving objects can be predicted. Mainly using MVs, this method does not fully utilized the information from compressed domain. In [10], a classifier based on support vector machine (SVM) is trained to determine if an image patch belongs to the foreground objects or background using HEVC syntax elements including MVs, coding unit (CU) size, etc. Only one frame’s syntax elements are used without considering temporal adjacent frame decoding information in this method. Moreover, both of them are static-camera-based. They can’t handle with moving camera situation, such as camera jitter. In this paper, a novel moving segmentation method in HEVC compressed domain based on logistic regression classification is proposed. The proposed method applies to camera jitter scenarios which is common for surveillance cameras installed in the outdoors where wind blows. Besides, both spatial and temporal information are fully utilized. Experiments show that the proposed method can achieve comparable performance with state-of-art pixel domain method and 16.78 times faster processing speed.

ABSTRACT. SFN deployments are the optimal network deployments for large areas demanding a common broadcast content for signal distribution uniformity and spectral resource consumption reasons. Nevertheless, they have been sometimes disregarded for not enabling simultaneous delivery of global and local contents. By using LDM the delivery of local contents in an SFN has been proved possible. The injection level impacts the final network configuration as well as the bitrate allocations to the Upper and Lower layer contents (global and local). The values will be discussed using a specific network deployment in Washington DC. The final paper will include other service areas where the terrain characteristics will have an impact on the final coverage.

ABSTRACT. This paper proposes a superposition decoding scheme based on raptor-like LDPC code for signaling protection in broadcasting system. The proposed scheme takes use of the fixed signaling bits to improve decoding performance significantly and lower the decoding threshold with a few frames sacrificed. Meanwhile, this paper also proposes a selection algorithm to select proper LDPC bits positions for fixed signaling which is better than random selection

ABSTRACT. See abstract attached.

ABSTRACT. This paper proposes a services pushing and caching algorithm in high-speed train (HST) by using converged wireless broadcasting and cellular networks (CWBCN). Services pushing and caching on the HST is an efficient way to improve the capacity of the network; and it can also lead to better user experience. In the proposed services pushing and caching model, the most popular services are delivered and cached on the vehicle relay station (VRS) of the train ahead the departure time. Then, the most popular services are broadcasted and cached on the User Equipments (UEs) after all the passengers are on the train; the less popular services are transmitted to the user by p2p mode by the relayed cellular network on the train. In order to maximize the network capacity in limited time slots, we transform the issue into the 0-1 Knapsack problem. Dynamic programming algorithm is adopted to solve it in polynomial time. As the passengers may get on or get off the train when pushing the most popular services, an information retransfer algorithm is also proposed when more intermediate stations are considered. Simulations show that the proposed algorithms can efficiently improve the capacity of the converged network.

ABSTRACT. Executing an accurate full-reference metric such as VQM can take minutes in an average computer for just one user. It is unfeasible to analyze all the videos received by users in an IPTV network for example consisting of 10.000 users using a single computer running the VQM metric. Therefore it would be required to use other type of systems or lightweight no-reference metrics in addition to the full-reference metric mentioned. Lightweight no-reference metrics can be used for discarding potential situations to evaluate because they are accurate enough for that task, and where more accuracy is required the full-reference metric VQM could be used. The main work in this paper is focused on determining how many situations/users can be analyzed using the VQM metric in a computer with good performance. The full-reference metric is applied on the transmitter using a method specified in the recommendation ITU BT.1789 “A method to reconstruct received video using transmission error information for packet video transmission”.

ABSTRACT. As the Internet spreads, new web applications come to light. One promising sector that is still in its infancy is supermedia applications. Supermedia applications process video, audio, haptic and other sensory data. Quite a few protocols have been suggested for the transfer of real-time supermedia data over the Internet. This paper outlines the related work on these protocols. It describes the Quality of Service requirements for supermedia applications that a network has to fulfill. Extensive simulations and experiments for the performance evaluation of transport protocols for real time transferring High Efficiency Video Coding streams with supermedia data are carried out. Complements, differences and relevancies between simulation and real world experiments are discussed. The metrics that are measured for the performance evaluation are the delay, the jitter, the throughput, the efficiency, the packet loss and the proposed by the authors, packet arrival deviation. The simulation tests reveal which protocols could be used for the transfer of real-time supermedia data with a High Efficiency Video Coding video stream.

ABSTRACT. In this paper, the problem of peak-to-average power ratio (PAPR) reduction of orthogonal frequency division multiplexing (OFDM) signals is investigated in the context of second generation of digital video broadcasting for terrestrial transmission (DVB-T2) and the American digital video broadcasting standard ATSC3.0. As a multicarrier modulation technique is characterized by a high PAPR of the transmitted signal and OFDM is prone to non-linear (NL) effects of power amplifiers. DVB-T2 and ATSC3.0 have adopted a gradient-based tone reservation (TR) PAPR reduction technique which is based on an iterative process where, at each iteration, a predefined kernel is used to reduce one peak in time domain. Recently, a new TR PAPR reduction technique termed individual carrier allocation for multiple peaks (ICMP), based on a novel kernel definition has been proposed. However, it suffers from latency issues in ATSC3.0 and higher modes of DVB-T2. So, we propose another novel TR technique, grouped ICMP (GICMP). The simulation results show that GICMP offers better performance than the gradient-based DVB-T2 algorithm. Furthermore, it not only yields the same performance as ICMP but also has less latency.

ABSTRACT. Current broadcast systems make use of either single frequency networks (SFNs) or multifrequency networks (MFNs) to distribute content. However, neither of the topologies is efficient to distribute both global and local content at the same time. While SFNs are ideal for delivering global content in a frequency- and power-efficient way, MFNs are well-suited to deliver local content to the desired reception areas. This paper describes a technique to allow local service insertion (LSI) in Digital Audio Broadcasting (DAB) SFNs. DAB has been in use for several years around the world. Considerations towards LSI have been made before. However, those considerations focus on the transmission modes (TMs) II and III, which will not be supported in future revisions of the DAB standard. One obstacle for LSI in DAB is the influence of differential modulation. The technique presented in this paper uses an artificial phase reference symbol consistent with differential modulation. Furthermore, this approach is compatible with TM I. The proposed LSI technique is described in this paper and analyzed in terms of implementation issues and performance evaluation with respect to the introduced interference.

ABSTRACT. We put up forward a high-quality frame interpolation framework and its real-time implementation on heterogeneous computing system.To obtain a truthful optical flow field, the framework takes use of recent progresses in correspondence estimation including randomized search strategy and edge-preserved flow interpolation, and adapts them well into the system through a parallelized design.Firstly, dense optical flow estimation by PatchMatch algorithm is improved by the proposed propagation scheme of good matches inside divided pixel batches along with a new message passing phase between them.Then, the parallelization of this framework is two-fold: on device-level, different tasks such as flow estimation and edge detection are simultaneously executed on CPU and GPU, while on data-level, the interpolation of flow is performed in parallel for different pixels.Finally, with interleaved flow fields in forward and backward direction, multi-hypothesis based motion compensated frame interpolation will produce high-quality videos at low computation cost. Experiments on a 16-core CPU and GTX Titan 980 Ti GPU system achieve real-time frame interpolation of 1920x1080 videos at 60Hz. Objective and subjective comparisons demonstrate that our method has obtained state-of-the-art quality.

Frame interpolation has been an open question in video processing. It aims to compensate additional intermediate frames between the original ones so as to increase the smoothness of motions. It has been widely studied in frame rate up conversion [1,2] or video error concealment in transmission over networks [3]. Recent real-time cloud gaming application has further aggravated the urgency of this technology [4], because it is discovered that frame rate and image quality are of the most key factors affecting the service quality of cloud gaming. Other areas such as 3D imaging, augmented/virtual reality will also benefit from higher-frame-rate contents which can provide better sense of immersion.

However, the pixel generation of an entire frame with modern high resolution such as HD (High Definition) or UHD (Ultra HD) is inherently computationally expensive. To reduce the complexity, classical methods choose block based algorithms in both motion estimation and motioncompensated frame interpolation [5,6]. It eventually leads to a quality loss of the generated frames such as blocking and smearing effects. According to the previous works [1,5], highquality frame interpolation requires both accurate motion flow and motion compensated interpolation. Moreover, considering of the computation complexity, the parallelization possibility in an algorithm of both process is also crucial for real-time applications.

Our method is mainly inspired by recent progresses achieved by PatchMatch [7] which is a randomized correspondence matching algorithm. PatchMatch is to quickly find approximate nearest neighboring matches between two images. The key insight driving the algorithm is that some patch matches can be found via random sampling, and the natural coherence in the image allows a propagation of such matches to surrounding areas. This pioneering work has sparked many optical flow algorithms such as CPM [8], SPM-BP [9] and EPPM [10]. Particularly, in CPM [8], a coarse-to-fine scheme with constrained random search radius between adjacent levels of the hierarchical architecture is proposed. On the other side, based on an observation that the correspondence is often spatially smooth with discontinuities aligned with object edges, SPM-BP [9] shares the matches inside a superpixel set by utilizing SLIC superpixel segmentation [11] and finally achieves a speedup of convergence of the correspondence field.

The propagation for PatchMatch is of significant importance on good matches to a fast convergence. But the dependency chain between neighboring pixels restricts the acceleration of it by parallel computing. Thus, in this paper, we propose a new PatchMatch based optical flow estimation algorithm as well as a highly parallelized implementation while maintaining the good performance of original methods. Specifically, we divide images into pixel batches which can propagate good matches in parallel and introduce a new message passing phase to guarantee local good matches of a batch to be propagated to others. Based on this improved PatchMatch algorithm, we develop a frame interpolation framework which still keeps a parallelized architecture. Different tasks such as flow estimation and the edge detection for flow interpolation are simultaneously executed on GPU and CPU, which is a device-level parallelization. Since the flow interpolation and frame interpolation are uncorrelated on the data-level, they are fully parallelized too.

ABSTRACT. High mobility along with multipath propagation in mobile reception results in a doubly selective channel model that exhibits selectivity in both the time and frequency domain. The estimation and subsequent compensation of such a channel is a non-trivial task. The DVB-T2 standard defines the use of Orthogonal Frequency- Division Multiplexing (OFDM) in the physical layer that along with many inherent advantages, also helps in reducing Inter-Symbol Interference (ISI) by the use of a Cyclic Prefix (CP). However, Doppler shift caused due to high mobility in these systems result in a loss of orthogonality between the sub-carriers. Thus it is appropriate to estimate Doppler shift in the time domain using time-domain pilots. Although the DVB-T2 standard specifies different pilot patterns in the frequency domain, a pilot/reference pattern is not available in the time domain. In this paper, we propose time-domain pilots in DVB-T2, a novel approach that aids in the estimation of a doubly selective channel by any suitable channel estimation algorithm. The proposed technique retains the structure of the transmitter chain and is computationally feasible from an implementation standpoint. Moreover, it enables a precise estimate of the channel that would improve the performance of the DVB-T2 receiver in challenging channel conditions.

ABSTRACT. A Quick Response (QR) code is a two-dimensional bar-code, which is quite popular due to its excellent storage capacity and error resilience. Generally, QR codes are widely used to store text information such as URL links, geographical coordinates, name cards, inventory information, authorship, etc. To reach the current limit of QR codes, in this paper, we would like to propose an innovative multimedia archiving technique, which is built upon the advanced signal processing scheme to tackle the multiple QR codes all at once. The multimedia data often include texts, images, audio data, special-purpose codes, etc. The recently proposed software-defined multiplexing code (SDMC) can be applied to combine all of them even though each type of data would have the individual data format different from others. Our proposed new archiving technique involves two phases, namely multimedia-amalgamation (MA) and multimedia-detachment (MD). In the MA phase, the multimedia data, regardless of their data format, can be converted to the binary streams; then the SDMC will be employed to aggregate them together in a longer binary stream; such an ultimate long binary stream will be converted to the QR codes. The QR codes, protected by the inherent error-correction mechanism, will thus be placed on a sheet or multiple sheets with a uniform spacing. The sheet(s) containing multiple QR codes can thus be archived in soft (PDF) or hard (print-out) copies. In the MD (recovery) phase, one can scan these PDF files and employ our designed signal processing algorithms to separate the individual QR codes; then the corresponding QR decoder can convert these QR codes back to the original binary stream. Finally the unstuffing algorithm in the SDMC can detach the individual data from the long binary stream composed by the multimedia mixture.

ABSTRACT. Research on the compatibility between digital sound-broadcasting service and aeronautical radio navigation is very important to promote the development and popularization of FM-CDR in China. This paper analyzed the possible interference on the aeronautical radio navigation ILS/VOR caused by digital sound-broadcasting signal FM-CDR, and based on relevant compatibility research about the sound-broadcasting technologies such as FM, DRM and HDRadio, this paper presented measurement program to be used for actual testing. Through device selection, platform construction, preliminary test, data collecting and statistical analysis, the measurement program is confirmed to be reasonable and this paper summarized the compatibility conclusion between FM-CDR and ILS/VOR under the case that the frequency of the received signal and interfering signal is identical.

ABSTRACT. Digital television has brought many benefits to viewers: improved audio and video quality, better coverage, more signals and programming guide, among others. Interactivity through television is one of the biggest promises that brought DTV. However, a major deployment of interactivity through television is still pending. Possible explanations are that receivers must have additional costs by adhering to standards, or that interactivity did not enabled killer applications. More than 10 years ago, standards were developed that contemplate a DTV receiver receiving TV through traditional antenna system and simultaneously connected to broadband Internet, receiving contents by both ways, e.g. TV-Anytime (https://tech.ebu.ch/tvanytime/) [1]. These contents can be traditional video and audio, but also information, executable programs or other data. Nowadays, there are new standards with enhanced capabilities, which are in different stages of development and in general are in an emerging deployment. Among them are: - Hybricast [2] - Hybrid Broadcast Broadband TV 2.0 (HbbTV 2.0). [3] - ATSC 3.0 [4] - Ginga [5]

Other systems allow the existence of a connected TV. Such as the current "Smart TVs", which have limitations and restrictions due to commercial agreements with manufacturers. Also, there are televisions or STB using a standard operating system such as Linux or Android. These bring the promise of allowing the installation of applications from varied origins. Many voices in the world of television affirm that true television interactivity will happen through the so called "second screens", tablets or smartphones available to the viewer while watching TV. In fact, some statistics show that 40% of Americans use their second screen while watching TV. The recently published ETSI standard HbbTV 2.0 [3] is a promising enabling technology. Moreover, there is a new protocol called DIAL [6] that allow second-screen devices to discover and launch apps on first-screen devices. It is being rapidly employed due to the popularity of its supporters (Netflix and Youtube). DIAL is also included in HbbTV 2.0. Meanwhile, augmented reality has long existed [7]. Nowadays, it enjoys a growing popularity due to the multiplication of professional and home use services and devices that support it [8]. In fact, analysts say these new kind of services can be the killer application for the future of broadcast television [9]. Most recently, there have been proposals for Interactive Digital Television with augmented reality applications. These are experiences in which an application runs on the TV set, and other related applications runs on a second screen device, or even several second screens if there are multiple viewers simultaneously watching the same TV. Examples include educational games for groups of children, experiences in which the viewer can join a band during a performance, applications in where objects are displayed on the TV and evolve depending on how the second screen is handled... Published works mention issues on synchronization, implementation of protocols, standards, hardware and others. [10] This paper describes an ongoing project to develop a hybrid TV system that allows the execution of augmented reality applications. The system seeks to minimize the capabilities required in the hybrid receiver or set top box. We have developed a test platform, described in [11] and [12] that includes an ISDB-T receiver capable of executing HbbTV 2.0 applications standalone or in conjunction with second screens.

Figure 1. General overview of the system [8].

Figure 1 sketches the structure of the test platform. The constituent elements are: - A receiver running on a PC, which has an ISDB-T tuner, and is controlled by an HbbTV extension of Firefox. It has Internet connection, HTML 5 support, capability of overprinting HTML layer above the video and audio program, and capability to install and run applications synchronized and related to the broadcast content (the Receiver Application, “RXapp”). - The broadcaster needs to provide synchronization between the broadcast and the related content (RXapp). This can be performed using Timed External Media Information (TEMI) or other alternatives (that will be described in the full paper). - Second screen(s) with an app installed. - An application provider server that controls each receiver application and its associated second screen applications. This site also receives the signaling from the broadcasted Transport Stream and accordingly commands the apps that the receiver and the second screens display. The applications developed so far involve many different aspects and use cases that the new standard HbbTV 2.0 allows. 1. An app to answer trivia questions through the second screen, without the need of an HbbTV terminal. However, HbbTV synchronization tools are employed in order to ensure proper functioning of the service. The application running is shown on Figure 2.

Figure 2. App with trivia in second screen. Using HbbTV system but without HbbTV terminal. 2. Apps running synchronized between a TV set with HbbTV 2.0 capabilities and a second screen device. These apps allows a user to share with other TV viewers, on the TV screen, complementary content that he may find on the Internet with his second screen. This is shown in Figure 3.

Figure 3. Apps at main and second screen. Sharing of Internet info from second screen to the main screen. 3. Apps running on a TV set with HbbTV 2.0 capabilities and on second screen, that allows the publication of complementary content provided by the broadcaster, synchronized with the main content being broadcasted. This is shown in Figure 4.

Figure 4. App on main and second screen. Moving one 3D object (complementary content) in second screen published by broadcaster to be consumed simultaneously.

As far as we know, this is the first report of HbbTV 2.0 applications running in an ISDB-T transmission environment.

REFERENCES [1] Alberto Gil Solla, and Rafael G. Sotelo Bovino. TV-anytime: Paving the Way for Personalized TV. Springer, 2013. http://www.amazon.com/TV-Anytime-Paving-Way-Personalized-TV/dp/3642367674 [2] Integrated Broadcast-Broadband System Specification, IPTVFJ STD-0010 Version 2.0 [3] Hybrid Broadcast Broadband TV. ETSI TS 102 796 V1.3.1 (2015-10) [4] atsc.org/standards/atsc-3-0-standards/ [5] Digital terrestrial television – Data coding and transmission specification for digital broadcasting Part 5: Ginga-NCL for portable receivers – XML application language for application coding, ABNT NBR 15606-5, 04/09 [6] DIAL (DIscovery And Launch) http://www.dial-multiscreen.org/home [7] Azuma, Ronald T. "A survey of augmented reality." Presence 6.4 (1997): 355-385. [8] Kipper, G., & Rampolla, J. (2012). Augmented Reality: an emerging technologies guide to AR. Elsevier. [9] Cuthbertson A., Forget Reality TV, Virtual Reality TV is the future of broadcast television, http://www.ibtimes.co.uk/forget-reality-tv-virtual-reality-tv-future-broadcast-television-1489400 (checked online Dec 29th 2015) [10] Kim, J., Ha, J., Choi, B., Jeong, Y., & Hong, J. (2014). Design and Implementation for Interactive Augmented Broadcasting System. Broadcasting, IEEE Transactions on, 60(2), 217-226. [11] Jose Joskowicz, Rafael Sotelo, “Integrating ISDB-T Broadcast Signals with HbbTV Internet Broadband Services,” in 2016 8th IEEE Latin-American Conference on Communications (LATINCOM), October 2016 (in Press). [12] Rafael Sotelo, Jose Joskowicz, “ISDB-T and HbbTV Hybrid Receiver,” in IEEE International Conference on Consumer Electronics (ICCE), January 2017 (in Press).