ABSTRACT. This paper will propose a technology-oriented model for regulating Self Driving Vehicles (Hereinafter: SDVs). SDVs challenge current regulation built around the existence of a human driver controlling the vehicle and responsible for its operation, with a set of machine-built algorithms. These statistics-based algorithms control the vehicle aided by sophisticated sensors and information from the internet and other vehicles replacing the human eyes and ears. The lack of a human driver creates multiple simultaneous challenges in different legal fields – some more apparent than others. A comprehensive solution for liability is required to balance SDV prices with raising safety levels in an era without human error, efficiently allocating resources between manufacturers and car owners,. Privacy must be addressed as we maximize data gathering required for navigation while avoiding de-fact surveillance by the thousands of SDV mounted cameras which. The public must feel safe knowing that SDV algorithms will always act in their benefit and not programmed to save the rich or the interests of the manufacturers. Regulators should be assured that the legal mechanism will be stable enough for as long as possible even when the technology changes. To achieve a comprehensive solution, instead of regulating the end product, a technology-oriented model breaks down SDVs into their main technologies – the algorithms, the sensor array, the connectivity technologies and the mechanical vehicle. Analyzing each technology separately, this model efficiently and precisely maps the legal challenges and offers better suited legal solutions, long-lasting and easier to adapt than those based solely on regulating the end product.

ABSTRACT. Most of the research on autonomous vehicles (AVs) highlights the potential disruption that this technology might have on urban mobility. However, the discussion is usually not grounded in a social theoretical foundation. AVs impacts depend ultimately, on how the technology will be adopted by users (e.g. usage of private or shared AVs) and integrated into the socio-technical mobility system. This study aims to explore the social aspects of the transition towards SAVs and future usage models. The research combines mixed methods (qualitative- 5 focus groups, and quantitative- survey of 1000 participants) in order to explore the multidimensional aspects of AV adoption. First, a theoretical exploration illuminating individual and societal insights for SAVs transition was conducted, and then the hypothetical formation of a new form of travel based in the formation of travel communities (TC) in the AV era was tested. TC is defined here as a group of persons formed around individual preferences regarding social and travel characteristics with the aim of sharing a trip in a random AV. The empirical findings support the potential and importance of social considerations in decision-making process of mode choice such as the loss of flexibility, privacy and stranger concerns, and generational different perspectives towards mobility innovation regarding sharing the trip in the AVs era. The social considerations presented by the personalized individual preference (PIP) concept in addition to attitudes towards sharing in general present an extra social value for adopting a sharing form of transport. This approach could be also useful for existing transport sharing models. Also, planners and stakeholders might take into consideration the relevance of individual and societal aspects for the promotion of new forms of mobility based on sharing models.

ABSTRACT. Autonomous vehicle (AV) implementation as a key area of the future mobilities revolution presents an opportunity for cities to reset deep-seated problems related to automobility, yet conversely also risks exacerbating these very urban problems of car-oriented city planning, decline in public transport, environmental burdens, and produce new, or aggravate existing, patterns of injustice.

The current process of AV introduction is being shaped by corporate players, interest groups and governments acting in an uncertain environment. It is unclear how much the technology can realistically achieve; to what purpose, how, and the extent to which to regulate the technology; and how to harness the technology for public interest, taking into consideration potentially diverse and even conflicting goals and values. Against this backdrop, public values (Bozeman, 2002, 2007) can be used to strategically and actively guide the direction of technological implementation that can be harnessed for cross-cutting public interest and socio-urban goals (Guo and Marietta, 2015).

From our interviews with AV professionals in Israel from different sectors and industries, we distil a range of public values. We examine how professionals formulate AV issues; how they position their arguments in interaction with other professionals; how they deal with conflicting interests and knowledge gaps.

We set up a conceptual framework of the Recognition-Consensus Model for public values through which we categorise four groups of values: values of high social accord, low social accord, high contention, and subordinate concerns. The analysis of these value groups can help to discern which are the high priority considerations, to identify problems with dominant narratives, to understand key areas of dissensus and to identify gaps in problem formulation and potential policy gaps.

ABSTRACT. The flexibility of shared automated vehicles may engender new transportation modes that will be a cross between the private and public modes available today. We analyze user preferences towards shared automated vehicle services, including ridesharing, car sharing and automated transit, using a hybrid choice model. We use a discrete distribution to capture taste heterogeneity of distinct latent classes. Latent variables, socio-demographics and travel habits inform latent class assignment estimated simultaneously with a discrete choice kernel. Respondents chose their preferred mode for going to work in a set of stated preference choice tasks, based on the attributes of their current commutes using a Bayesian D-Efficient design. Users can be segmented into two latent classes based on latent factors that capture personal time style orientation and public transit dislike, based on a multiple indicators multiple causes confirmatory factor analysis model. Effects of trip cost, travel times and seating designation in a shared ride, were estimated for the two classes. Users who neither like transit nor ridesharing with strangers, are less likely to choose a shared ride if their designated seat is the middle seat, and overall less likely to choose automated transit. Individuals who have more organized time styles demonstrate higher marginal sensitivity to travel times and costs and are more likely to choose automated transit. Value of time analysis reveals that wait time of services that offer a convenient home pickup is valued lower than in-vehicle time. The effects of multitasking while traveling and travel habits are also modelled and discussed.

ABSTRACT. An external human-machine interface (e-HMI) that displays a message indicating the Fully Autonomous Vehicle’s (FAV) intent can improve communication between pedestrians and FAVs. However, how the form of the message, the crossing environment, and the number of FAVs affect pedestrians’ road crossing decisions is yet unclear. This study examines pedestrian crossing decision and message comprehension when one or two FAVs are present at the scene. In an online experiment, 119 participants, 23-68 years old, conducted saw 24 urban road crossing scenes. Scenes, generated in a synthetic environment, differed in message type (status/advice), display color (red/green), and crossing complexities (Vehicle’s distance from crossing point, number of FAVs - one / two). The e-HMI was identical when two vehicles were present. Participants crossing decisions and answers regarding the messages meanings were collected. In addition, participants filled the Attentional Control Scale (ACS) Questionnaire. Results showed no significant difference between pedestrian decisions when one or two FAVs were present in the scene. Overall, 72% of the crossing decisions were consistent with the FAV message (compatible decision). The lowest compatibility (43%) was when the FAV was close to the crossing point, and the display was green. Advice messages led to more compatible decisions (76%) compared to status messages (68%). Results confirmed previous studies indicating the message type and color display effects on crossing decision. Further, the number of FAV did not influence pedestrians' decisions. Also, in situations that could be interpreted as dangerous, pedestrians' decisions relied mostly on the FAV’s distance from the crossing point rather than on its recommendation. The study's main limitation is the identity of the display between the two FAVs (100% agreement). To extend findings regarding the effect of multiple FAVs on pedestrian decisions, it is necessary to further examine parallel contradictory messages in different crossing complexities and dynamic crossing situations.

ABSTRACT. Autonomous (automated) vehicles, whose development began about a decade ago, in many western countries, are already traveling the world's roads, some on sterile-experimental sites and others on freeways. Regulatory decisions on this issue are slow, and it can now be felt that the lack of regulation (of a positive kind) will stop companies' ability to inculcate developments on state roads. The impact of smart transportation on the Middle East and the underdeveloped countries is going to be enormous (direct and indirect), such as: the decline in the use of polluting fuels; Or full dependence on Western technology. Regulation requires many ethical decisions, including the decision on the organization that will make the necessary decisions. The significance is clear - vehicles will not be systematically adapted to the behavior of the Middle East population and non-western countries, should such a situation be allowed? Does the fact that autonomous vehicles, which can save 95% of road accidents, not be adapted to such a large population with the highest population density and birth rate? Will this not lead to a kind of popular uprising in the future?

ABSTRACT. Israel's drylands are being populated, the air is filled with planes, and our eyes are looking toward the horizon, the sea. Until today, suggestions for alternative airports facing Israel have been rejected due to ecological and financial reasons. Project Anthazoa offers an advanced alternative in the form of a fragmented airport. In the project, one can see the terminal receiving passengers in Tel Aviv, and the floating marine structure assembled in the middle of the sea, who's underwater structure provides a platform for marine wildlife to grow life on. Anthozoa Airport is located 15 kilometers from the shores of Tel Aviv. The structure of the Anthozoa Airport provides a breeding ground for corals and creates a marine habitat. Reef for the soft sea animals. The journey begins in Tel Aviv. Where the passengers arrive for check-in, security checks, and take the train to the sea.The reception terminal merges with the urban atmosphere and functions as a cultural and commercial center. The airport will absorb millions of passengers a year from Israel and around the world. Thanks to the dismantled structure, the absorption and transportation of passengers is very streamlined. The land terminal affects the surrounding city, and strengthens the sense of connection of the city to the aviation artery.

ABSTRACT. Israel takes part in the global trend of adopting a sustainable transportation policy. However, it seems that policymakers in Israel tend to rely on insufficient data when design transport policy. Moreover, the governmental database does not stand in global standards, making it challenging to evaluate sustainable transportation progress. Along with the importance of relying on data, insufficient work is done by the government to create a relevant and useful database just by analyzing and integrating existing databases.

Our goal in this study is to emphasize the need to establish a governmental transportation database, which integrates aspects of the service level from the travelers’ perspective.

First, we review existing national transportation databases of several countries and the UN Indicators for the Promotion of Sustainable transportation. Then, we present the existing Israeli databases, maintained by the Ministry of Transportation and the Central Bureau of Statistics. To demonstrate the importance of analyzing and integrating existing databases in the decision-making process, we conducted a study examining Israel Railways stations' accessibility. We utilize three data sources: GTFS file, Israel Railways data, and a survey we conducted (N570=). The study integrates data such as means of transport used to get to the train station, traveler satisfaction, volume of buses, trains, and passengers at the train station during the day, and the availability of a parking lot for various transportation means. By focusing on one station at a time, we observed the specific accessibility problems associated with each of them. We conclude by suggesting recommendations to the government regarding establishing a national database for public transportation that stand in global standards.

ABSTRACT. In recent years, parking, or the lack of it, has become a major concern for both the citizens and the municipality of highly dense cities. To solve that problem, several companies has developed automated parking lot (APL), where the driver enter with her car to an entry/exit room, from there a series of robots takes the car to a parking space within the parking lot and retrieve it on demand. This type of parking lot has a huge advantage in space efficiency over the conventional parking lot design. The type APL we are interested in uses shuttles to carry the cars within the parking lot. The shuttles moves on a single horizontal rail in each floor and two elevators, located at both ends of the rail, are used to move the shuttles between floors. This design generates several operational issues, mainly, the allocation of tasks to shuttles and the routing of the shuttles. Both issues are directly influenced by the allocation of parking spots to cars, an issue that is common to all APL types. To solve the shuttle routing issue, some companies uses a routing method based on circular movement. Where, one elevator is used to take the shuttles from the entry floor to the parking floors, while the second is used to take the shuttles from the parking floors to the entry floor. The shuttle moves in one direction in the entry floor and in the opposite direction in the parking floor. Under this operating method, the utilization of the elevators is at best 50% and they are the bottle neck of the process. Using queueing theory, we proven that this method is sub-optimal in some cases. We are now working on new routing methods to different demand scenarios and a simulation to put them to the test.

ABSTRACT. When consumers buy energy-efficient cars, they tend to drive more. The economic literature defines this phenomenon as a ‘rebound-effect’, estimating that 20-40% of the potential fuel savings associated with cars’ improved energy-efficiency are lost to increased driving. However, whether the rebound is constant over time remains unclear. This question is substantial because in 2019 policies that encourage consumers to purchase energy-efficient cars covered nine out of every ten new cars sold worldwide. In 2009, the Israeli government implemented a policy that incentivized the purchase of energy-efficient cars. In previous work, we estimated a rebound-effect of about 40% around this period. The present study examines the presence, or lack thereof, of a long-term rebound. Specifically, we examine whether consumers increase (decreased) their car usage in the long-term, thereby contributing to a larger (smaller) rebound. We utilize a decade-long household survey (CBS, 2007-2016), including car usage, car characteristics, and socio-demographic data. We used a simultaneous regression model to address the endogeneity of households choosing a car with a specific energy-efficiency level considering their expected car usage. To account for potential non-stability in the causal relationship over time, we adopted a ‘rolling-window’ technique. Our findings indicate a rebound-effect of 40% a year after the introduction of the policy. Whereas it remains relatively steady in subsequent periods, it later gradually intensifies until it reaches 98%. This unique finding suggests that virtually all the potential energy savings due to energy-efficiency improvements were lost to increased driving. We found that a policy incentivizing energy-efficiency may increase the rebound-effect in the years following its implementation. We postulate that the reason for this phenomenon is a learning process that consumers experience, during which they become increasingly aware of usage costs over time. This increased awareness may encourage energy-efficient car owners to drive more, thereby saving considerably less energy.

ABSTRACT. We are an international team of designers, artists and creative technologists that challenges traditional boundaries in science, technology, and society. This unique collaboration between members of the DLX Design Lab, Institute of Industrial Science, at the University of Tokyo, and students and lecturers at Bezalel Academy of Arts and Design, Jerusalem started in November 2019. Since then, we have been working on a project dedicated to the challenges presented by autonomous vehicles (AV’s)entry into the public space.

Interests and Motivation

The Design Bezalel Mobility team, has been conceptualising a new set of gestures, imagining a new rapport between AV’s and humans in the near-future. Our proposal relies on existing communication patterns, patterns we consider as a language to be adopted in this field; this language could instill trust in humans who share the public space with AV’s, be they passengers, pedestrians, drivers, or other road users. Beyond ensuring that safety is maintained, we aspire to counter alienation by taking an holistic approach to these forms of communications, considering aspects such as human attention and road conditions. The communication patterns we’re developing rely on technologies present in today’s street, but also on sensors, cameras and face recognition technology for improving the information flow between humans and AV’s. We would like to reimagine the communication with the AV, employing potential modalities such as light, motion and sound. It is our goal to create a playful and engaging multimodal experience, one that centers on human needs and to stretch the boundaries of non-verbal, cross-cultural communication which we deem essential for the upcoming future.

ABSTRACT. We propose an information visualization tool to help urban transportation planners to analyze the public transit (PT) network from real-world data. Considering passenger volumes, together with bus route volumes and speed, the Israeli Ministry of Transportation published in 2016 guidelines explaining the circumstances under which a road link could qualify for a bus lane allocation.

We describe the design and development Bussleneck which allowed users to observe and analyze MoT guidelines in real-world use cases. Users can analyze the underlying metrics in-depth, while the tool provides confidence in the data processing, metrics calculation and presentation. Different data sources: bus schedule, historical real-time aggregated bus movements data, and passenger counts at stops are analyzed and processed into visual elements for analysis.

Detailed and thorough design of user tasks and visual mapping was conducted to produce visual means for analysis. The tool includes an interactive histogram, a scatter-plot and a geographical map to help users understand the metrics of bus delays, bus speed and passenger counts on each and every public transit segment in the cities of Beer Sheva and Tel Aviv. An expert user study will be conducted in order to evaluate the tool’s usability

The tool assists in promoting a real-world data-driven bus lane planning, bringing theory into realization. Bus lanes are crucial in delivering consistent and reliable public transit service. If we wish to encourage people to use more PT, we must deliver the best service we can where reliability is a key factor. Our tool can help planners in the process of bus lane allocation by providing them justifying evidence. Moreover, since one of the major obstacles of promoting bus priority means is objections by the public, this tool can also assist authorities in explaining major bus lane projects at the expense of general traffic lanes.

ABSTRACT. Public transport is a vital component of the transportation network and is supposed to provide transport services to citizens in an efficient manner. For public transport to be efficient, it must be adaptive and able to address demographic growth and changes in land use. The accessibility measure can be defined as a function of several destinations that are connected by a public transport line. Moreover, the distance a person must cross to reach the bus stop nearest his/her home can also be associated with the accessibility measure. The important accessibility aspect is the access to the geographic space from a specific location within a specific travel or riding time. The physical access to transportation, i.e. the distance of a potential passenger from the transit stop. In most cases, a 5-minute walk is considered a reasonable time for reaching a stop. In terms of physical distance – the acceptable standard set for urban areas is about 400 meters from stop to stop. The study found that there is a relationship between station location and connectivity and frequency of public transportation. As the distance from the center of the settlement increases, public transport accessibility options will even increase. If the proportion of people who can use complementary transportation ( as bicycles, scooters, shared cars, etc.) as a means of getting to the buses stations will be increased, it is expected a higher probability of using the public transit system. On this study we used transit database that includes line routes, bus stops, and schedules in GTFS format, which enabled to locate bus stops. Our research proposes a novel approach to analyze buses stops locations in relation to rural localities by using a combination of data science methods, calculating network-based distances including GIS-based analysis and statistical modeling.

ABSTRACT. With the increase in population densities and environmental awareness, public transport has become an important part of urban existence. Large quantities of transportation data are generated, and today, a standardized method to understand travel habits comprises mining data from smart card use. Public transport datasets may lack data integrity; boarding stop information may be missing due to either imperfect acquirement processes or inadequate reporting. Large quantities of observations may be missing from the smart card database. We developed a machine (supervised) learning method to impute missing boarding stops based on ordinal classification. A new metric, Pareto Accuracy, is suggested to evaluate algorithms where classes have an ordinal nature. Results are based on a case study in the city of Beer Sheva during which one month of data was collected. We show that our proposed method significantly outperforms current imputation methods and can improve the accuracy and usefulness of large-scale transportation data. We applied transfer learning and used our model in another city where we obtained similar results.

ABSTRACT. Dynamic Wireless Power Transfer (DWPT) technology offers many advantages for the future Electric Vehicle (EV) economy. Transmission coils, embedded in the road, transmit energy wirelessly to receiving coils mounted beneath the chase of the EV passing by. In order to achieve high energy transfer-efficiency, DWPT operates under resonance conditions and highly depends on the magnetic coupling, k, between the receiving and the transmitting circuits. Utilizing inserted capacitors in these circuits, it is a common practice to fix the operating frequency of the circuits at the self-resonance frequency of the system. This work investigates the dynamics of the magnetic coupling during the pass of the receiving coils setup over the transmission coils array. The position dependent k is extracted and shown to affect the energy transfer efficiency. To improve the efficiency of DWPT, we propose to vary the operating frequency and make it position-dependent. To still remain under resonance conditions, the capacitance in circuits has to vary accordingly.

ABSTRACT. As some of the researchers suggest that Covid19 probably will stay with us in the next few years. We must learn and adapt and live with that. Our research focuses on the development of an autonomous robotic charging system for electric vehicles. These days the world is moving towards fully autonomous vehicles, while more and more car companies are developing electric vehicles. In order to achieve full autonomy of vehicles and eliminate the interactions and possibilities that someone may get infected, a technology of autonomous charging must also be provided. The research includes developing a snake-like robot and special end-effector in a way that allows charging without any human involvement. This research is based on previous research performed in our lab, in which the robotic system was developed (for other uses). The robotic system has a novel design that includes a hyper redundant articulated robot where the electric cable is passed through the hollow robot links. Each link is separately controlled so there is no coupling so ever between them. The joints are controlled by pulling/releasing strings from the motion unit at the base. A specially designed end-effector for connecting to the charging socket. In addition to mechanical development, we also develop an AI algorithm for detecting the vehicle, charging cover, and socket using computer vision. This will include the control algorithm for directing the robotic snake and the end-effector to the vehicle socket, open it, and charge. All done autonomously without human involvement to reduce the possibility of infection.

ABSTRACT. Autonomous vehicles and motion planning are taking huge steps towards a new era for self-driving cars, planes, and more. With respect to autonomous racing cars, there are some important characteristics that need to be considered for the motion-planning to be useful, precise, and importantly – safe. Those characteristics, of the forces that applying to the car, are taking place in many fields of knowledge, such as algorithms development and implementation, structure and materials, simulation and experiments, etc. This research reviews and focuses on the aerodynamic characteristics that heavily affect the car's mobility, while we study them and try to create a reliable dynamic model for the autonomous racing car to self-drive and compete.

The main upcoming goal of this research is the international "Indy500" head-to-head racing cars competition, which Ariel University participates in. the competition's objective is to make the automation of vehicles better through the collaboration of world-wide universities that are a part of this head-to-head battle as well.

The research plan includes wind-tunnel experiments and simulations for the validation and verification of the aerodynamic coefficients and effects, different optimization methods all combined to motion-planning algorithms, and hopefully eventually in the "Indy 500 Challenge".

ABSTRACT. The transportation system is one of the largest and most important infrastructures in modern society. Modern transportation networks must satisfy high mobility demands, but at the same time, must do it with the lowest possible energy consumption and CO2 emissions. The sustainable transportation systems are based on three pillars, economic development, environmental stewardship, and social equity. One of the principles set for finding the balance between these pillars is to increase the use of renewable energy sources. Energy harvesting is a prominent research topic with many promising applications in transportation infrastructures and bridges. This work is devoted to the development of an efficient electrostatic generator for energy harvesting from mechanical vibrations. Analytical calculations are performed to evaluate the amount of energy generated. The energy generation principle is explained as “energy is invested into the capacitor at maximum capacitance and carried off at minimum capacitance.” A specific capacitor is designed for this purpose; one surface of each capacitor plate is coated with the oxide layer. The electrolyte-soaked porous membrane was placed between both the electrodes. This method significantly reduces the effect of the air gap, as a result, the efficiency of the generator increases. Preliminary experiments showed that electrolyte-soaked membrane increases capacitance almost 1000 times, which is a significant change for energy generation. More importantly, the initial capacitance of the proposed design can be achieved up to 1000 µF on the appropriate choice of dimensions. In this condition, the proposed electrostatic generator can provide energy of 50 J at the initial voltage of 10 V per cycle. If this proposed device is operated in the low-frequency regime of 10-100 Hz, it is possible to get electrical power of 2-25 W.

ABSTRACT. Semantic perception is the process of acquiring and building knowledge of both the geometric and the semantic structure of the environment. It allows an autonomous robot (e.g., vehicle) to "understand" the world in human terms, creating common language for a variety of tasks, and enabling high-level reasoning and decision-making. However, as semantic information is categorical in nature, probabilistic reasoning about semantics gives rise to mixture models with the number of components combinatorial in the number of candidate categories, which can in general only be tractable for real-time operation under approximations. Further, treating a finite number of pre-determined discrete categories inherently limits the information extracted out of the raw observations (e.g. images) which in general contain much richer information. We propose a different approach, whereby the semantic information attached to objects (or generally, scenes) is represented as continuous vectors in a latent space induced by a learned predictive observation model. The proposed observation model relates spatial changes in semantic measurements of an object to the latent object representation, allowing for joint inference of geometry and semantics free of discrete variables by maintaining a posterior over robot trajectory, geometric object and environment properties, and the learned object latent semantic representation.Being conditioned on a continuous representation rather than a discrete category allows such a model to be learned without requiring prior knowledge of candidate object categories. Efficient continuous inference under our approach rather than mixed-continuous and discrete inference which is normally required make our approach better suited for real-time operation (e.g. in autonomous vehicles), while relaxed requirements on ground-truth classification allow for easier adaptation of the model to new data.

ABSTRACT. This research proposes a novel combination of machine learning techniques and discrete choice models for route choice modeling. The data-driven choice set generation method identifies routes characteristics by clustering on normalized route characteristic attributes, and implicitly generates the choice set by sampling route characteristic attributes from the clusters. The proposed data-driven alternative sampling approach is inspired by the labeling methods. However, instead of artificially defining the labels and explicitly searching for alternative routes in the network with these labels, we sample the normalized route characteristic attributes to describe the alternative routes. Our approach tries to capture individuals’ interpretation on different routes by differentiating routes with their characteristics, rather than their explicit attribute values.

Important features are selected by random forests for route choice model development. The feature selection step not only reduces the computational power required for model estimation, but also helps reduce the risk of overfitting.

With the selected features, the methodological-iterative (MI) approach is applied to specify the utility functions and to find significant explanatory variables automatically and obtain a model with high goodness-of-fit and prediction power. The MI method can be seen as a two-level backward elimination procedure, in which the upper level checks all selected, and the lower level performs coefficient combination and elimination.

Results show that the proposed data-driven method produces a discrete route choice model not only with strong explanatory power, but also with high prediction accuracy compared to models estimated with conventional choice set generation methods.

ABSTRACT. We present a model for policy optimization of mobile personnel's online booking over a multiday horizon, with a different cutoff for each day. The system interacts with each customer, wishing to place a service request with the goal of maximizing the expected ratio of accepted requests in the long term. The demand forecast and user choice model are inputs of the policy optimization problem. The interactions with the customers are performed in a single step: The system offers an assortment of time slots during the service horizon, and the user either chooses one of them or abandons the system. The model maintains a tentative routing and scheduling solution that is updated after the acceptance of each request. The assortment of time slots following each service request is constructed by maximizing the expected net gain from the assortment. The net gain is estimated using a Cobb-Douglass function of features that represent the system's current state in a concise way. The parameters of this function are fitted using a simulation framework. The proposed method is benchmarked based on randomly generated datasets in various demand scenarios and geographies. The method is shown to significantly outperform a more straightforward baseline policy that is commonly used.

ABSTRACT. We present an approach for multi-robot consistent distributed localization and semantic mapping in an unknown environment, considering scenarios with classification ambiguity, where objects’ visual appearance generally varies with viewpoint. Our approach addresses such a setting by maintaining a distributed posterior hybrid belief over continuous localization and discrete classification variables. In particular, we utilize a viewpoint-dependent classifier model to leverage the coupling between semantics and geometry. Moreover, our approach yields a consistent estimation of both continuous and discrete variables, with the latter being addressed for the first time, to the best of our knowledge. We evaluate the performance of our approach in a multi-robot semantic SLAM simulation and in a real-world experiment, demonstrating an increase in both classification and localization accuracy compared to maintaining a hybrid belief using local information only.

ABSTRACT. At the end of 2018, "Waymo" launched a Shared-Automated-Vehicle (SAV) service in Arizona, USA. Shortly, the above service may dominate the urban transportation landscape. With technological advancement, SAV is expected to improve air quality, road safety, and so on. Future SAV services will probably operate in a "Stop-to-Stop" configuration, where people are dropped/picked at bus stations akin to the “Bubble” service that recently began operating in Tel Aviv and “Tik Tok” in Haifa and Jerusalem.

To assess the degree of adaptation to SAV, we use a multi-agent, multi-modal spatial simulation transportation model, with the addition of a future SAV service (using the "Stop-to-Stop" configuration) in the Tel Aviv metropolitan area (TAMA) and the Jerusalem Metropolitan Area (JMA).

The simulation was built in the spatial resolution of a single agent using the open-source Multi-Agent Transport Simulation - MATSim. Based on this study's results, a fleet of 75,000 SAVs will suffice to serve the entire population in the TAMA with no other transportation modes, which resulted in reasonable waiting/travel times and a low percentage of rejected SAV trips (less than 1%).

When agents are allowed to change to other means of transportation in the TAMA with a smaller fleet of 25,000 SAVs (priced at the cost of regular public transport) – SAV constitutes about 40% of all trips, 18% of which is derived from private vehicles and another 22% from public transportation. Agents who are rejected to use SAVs are close to zero.

The operational consequences of the transition to SAVs from other transportation modes in the TAMA and JMA will be discussed at the conference.

ABSTRACT. Independent mobility is important for the individual’s quality of life. The older population in Israel grows and becomes more diverse and complex. The purpose of our study, partly funded by the Ministry of Science, is to characterize the movement patterns of the third age population and their mobility habits. The database on which the study is based represents a survey carried out by the Master Plan for Transport, Jerusalem. As part of the survey conducted in 2010, a GPS device was given to every tenant in the household, which kept its locations and movements for a full day. The respondents in the survey were then asked in an interview about their activities during that day. A total of 8,253 households participated in the survey, including 33,265 respondents, and 590,151 activities.

Data analyses were conducted at three levels: the activity level, the tour level, and the respondent level. At each level, different characteristics were examined including the purpose of the activity, the travel mode, and the distances. The statistical methods in the analyses included a variety of methods, like a Pearson Chi-squared test, a Kruskal Wallis test, a Kolmogorov-Smirnov test, post hoc tests, and a regression model, which contained various methods within it. The results of this study indicate some trends with age starting at age 60, such as a decrease in driving activities, a decrease in the total sum of distance traveled by a person, a shortening of tour durations, and more. The findings indicate different trends for ages 20-60 versus those aged 60-100, as well as an internal age distribution within the third age group for group 60-80 versus the 80-100 group. Therefore, the processes of transport systems design should handle the needs of the third age, thereby facilitating their mobility and, consequently, supporting their lifestyle in general.

ABSTRACT. Pavement Condition Index (PCI) is commonly used in Pavement Management Systems (PMS) for indicating the extent of the distresses on the pavement surface. PCI values are a function of distress type, severity, and density (measured as a percentage of the total pavement area). The PCI values can be calculated manually or with the assist of soft computing technologies. Deep learning techniques, especially Neural Networks (NNs), have become increasingly popular in modeling engineering problems. NN techniques have successfully modeled the performance of the in-service pavements, due to its efficiency in predicting and solving non-linear relationships and dealing with an uncertain large amount of data. However, a major shortcoming of PCI calculation, is that it needs raw data preprocessing to be used later for the calculation or prediction. Usually, this task is done manually by different experts, which are time and effort consuming, furthermore, it may cause a discrepancy between the results, leading to improper estimation for the pavement performance. Therefore, the main challenge is to provide smart methods that can automatically detect and classify the distresses and their features and so, prediction the pavement performance. Our research suggests implementing smart techniques, such as object detection models, in order to accomplish the above-described task and overcome the challenges. According to the literature, such smart technology has the outstanding advantage of being capable of taking unprocessed images as an input, extract any desired information from the image, such as: to detect, locate, and classify objects and theirs features.

ABSTRACT. The first step toward Smart & Innovative Transportation is the adoption of non-motorized transportation (NMT) solutions or “soft mobility”. This includes various ways of mobility based on human energy (Human Powered Mobility), such as walking, cycling, skating etc. To complete the picture of alternative transportation, it is customary to examine electric vehicles as well. NMT is local by nature, given the limitations of distance for travel and mobility. In addition, unlike motorized transportation infrastructure, local authorities have higher independence to lead and promote soft mobility policy. Therefore, it is interesting to examine NMT presence and the ways local authorities promote initiatives in this important field. Which municipalities and jurisdictions carry out such initiatives? What characterizes them? What type of NMT is preferred? Answers to these questions can provide us some practical tools for optimally choosing policy measures to promote smart and non-motorized transportation, depending on the characteristics of the authorities, their capabilities, and the local conditions. Local government initiatives data were collected through a survey conducted among 104 local, urban, and regional environmental units that examined their environmental activities in a variety of areas, particularly in the field of non-motorized transportation. The insights from the study suggest that small local authorities are still a step behind bigger municipalities. Encouraging local entrepreneurship there can open a new arena for expending NMT solutions. Additionally, designing platforms for NMT policy initiatives, tailored to the needs and habits of specific populations (ultraorthodox, Arab minority), can be a key to the development of the field. This presentation is part of a PhD research project that identifies and compares local environmental policy initiatives (LEPI), under two different institutional structures, namely, centralized Israel and the decentralized confederation of Switzerland, as an engine of global change. This section examines the emergence of local initiatives in Israel regarding NMT.

ABSTRACT. Our goal is to save lives during emergency situations which require evacuation of the population. Those situations can be natural such as flood, tsunami, and earthquake, or man-made such as terrorist attack or a nuclear power-plant accident. There are many transportation models for evacuation. These models based on historical data which is usually obtained from surveys. Today there is new technology that was not available at the times the models were created. This technology is an online social network, and it can be used instead of the historical data. The online data covers wider areas and is much cheaper to obtain than the surveys. One of the models that best fit the evacuation during an emergency period is called a dynamic traffic assignment. We suggest using online location-based social network data to improve this model by suggesting a personal evacuation recommendation system. The location-based social network data gives us the friendship strength of the users and their locations. The locations are received from check-in data and the friendship weight is calculated from the interaction between the users, for example, the amount of "likes", "mentioned", and personal messages. Since people prefer to evacuate to relative or friends' houses rather than a community shelter, our personal recommendation system will calculate the nearest best friend which is outside the affected area. We created a tool – "Modular Multi-Dimensional Tool for Emergency Evacuation" (MMDT) that can integrates all the data sources and once there is a disaster, the tool we developed quickly finds the users who are in the evacuation area and maps each of them with the nearest best friend who is outside of the affected area. At the end of the process, each user has a personal destination and a personal route recommendation.

ABSTRACT. Dockless micromobility sharing services have great potential to change transportation in cities. In comparison to docked bikesharing, dockless sharing can improve mobility in underserved regions, reach more people, and increase micromobility usage. However, dockless micromobility sharing services could undermine existing public bikesharing systems. In this study, I compared the usage of three micromobility modes in Austin, Texas: docked bikesharing, dockless e-bike sharing, and dockless e-scooter sharing. I spatially analyzed micromobility trip logs with respect to the built environment and sociodemographics. In addition, I used a random forest model to examine the influence of different factors on micromobility trips. Results show that micromobility usage in Austin increased due to the introduction of dockless micromobility sharing services, while bikesharing usage did not decrease. Shared e-scooters and e-bikes are mainly used in regions served by bikesharing, and they increase micromobility usage outside these regions. E-Bikes are mainly used for commuting, while e-scooters and bicycles are used for short utilitarian trips. My findings suggest that cities should embrace all micromobility modes to cover all trip purposes and better serve their residents.

ABSTRACT. Cruising for parking stems from local mismatches between the patterns of demand and supply for parking. We propose ParkSage, a set of spatially-explicit algorithms for establishing spatially heterogeneous parking prices that guarantee a predetermined level of occupancy rate over a city, and for evaluating the reduction in parking search time. We apply ParkSage for establishing overnight parking prices that guarantee 85% occupation in the Israeli city of Bat Yam. Pricing by street links ensures high parking availability everywhere in the city, but is inconvenient for drivers. Establishing prices by the large and heterogeneous city quarters results in local mismatch between demand and supply, the emergence of areas with fully occupied on-street parking and long search time for the drivers whose destinations are in these areas. We demonstrate that pricing by the medium sized Transportation Analysis Zones, which is easy enough for drivers to comprehend and abide by, is sufficient for eliminating cruising. The software for establishing and assessing performance parking prices is based on the standard municipal GIS layers of streets and parking lots.

ABSTRACT. UAVs/Drones are widely used nowadays for various uses (shipments, surveillance, etc). The rapid rise in their use at urban regions cause traffic congestions and increase the risks of air-collisions. There is a need for an air-traffic regulator. Unlike regular airports, centrally controlled by a tower - due to the large geographic area, the terrain, the number of potential vehicles and the various stakeholders' interests – an automated, decentralized and transparent system is needed.

We introduce "AirBay" – a blockchain platform system which allows trading (sell and resell) of temporal air-right passages. We propose dividing the urban spaces into 4D pixels (x,y,z,t) - each pixel will have a unique numeric code to identify its 'cube' location through space and time-frame. Only The sole pixel-owner will be able to operate its flying vessels through the pixel's airspace at the allocated time. Clients can buy adjacent pixels to create a trail between endpoints and even a 'sub-space' grouping several pixels.

Our research deals with the technical aspects of the "AirBay" system under the blockchain infrastructure: a) How to encode the pixels (how to efficiently represent the airspace) to allow fast lookup of a transaction containing them - to determine their owner and in the authorization process, b) Various blockchain implementation schemes (smart contract vs. Unspent transaction model), and c) Evaluate the performance of such system (Time to sell, block size), as a proof of concept.

ABSTRACT. Traffic congestion is increasing as more roads are being built and cars being bought. However, the resources to increase road capacity have limited space and budget. One solution to improve traffic congestions are High Occupancy Vehicle (HOV) lanes along the general-purpose (GP) lanes. Such lanes can be utilized either for public transport, or for public transport and vehicles with multiple passengers. Another solution is the extension to High-Occupancy/Toll (HOT) lanes which allow even more vehicles access by paying a toll.

The pricing strategy managing HOT lanes has a high impact on its success. Low pricing might cause congestion, since too many travelers will be willing to pay. However, high pricing might result in low utilization, as more vehicles can use them without disturbing the free-flow. Many pricing methods have been developed, and they can be divided into three categories: fixed price, pre-scheduled (by the time of day), and dynamic pricing adjusted to the real-time demand. The relationship between toll price and saved time (Value of Time - VOT) traveling on HOT lanes directly affects the number of vehicles that choose to use them. This factor should be considered in any pricing method applied.

This study presents a generic procedure for finding pre-scheduled prices to optimize certain objectives, and shows a comparison to various existing pricing methods. The research follows real traffic data from the I-35E Minnesota express lanes (MnPASS). We find that not only do pre-scheduled prices have the advantage that the prices can be advertised so that travelers can schedule their departure time in advance, but the results show that using our pricing algorithm, the pre-scheduled pricing method provides satisfying performance, and in some measures can be even better than the dynamic algorithms. For example, maximizing operators' revenue, or minimizing HOT lane travel time.

ABSTRACT. Public Transport (PT) plays a major role in passenger mobility and contributes for sustainable transportation but must provide a continuous accessibility and connectivity for passengers, otherwise these advantages cannot be achieved. PT reliability is considered a major obstacle to market share growth. Current solutions to reliability are addressing mostly the travel time reliability (priority lanes, traffic signal priority). Dwell time reliability has not yet been addressed aside from increasing the use of smart cards, which indeed contribute to reduced boarding and alighting times variability. However, in-vehicle crowdedness causes delays and as a consequence increase dwell time variability. This work proposes a monetary approach, that dynamically changes the fare based on the in-vehicle crowdedness level. Similar to congestion pricing the ticket price is changed according to the crowdedness level. Doing so will enable to shift passenger from boarding the over-crowded vehicle, while compensating them due to the additional waiting time. On the other hand, passengers unwilling to wait might have to pay a penalty if the additional waiting time is resealable. In order to assess the attitude of passengers towards a dynamic fare model, a stated preferences questionnaire was developed in order to assess the attributes affecting the choice not to board an over-crowded vehicle. Based on panel data and the xtlogit model it was revealed that the higher the waiting time, the lower is the willingness to board the next vehicle. On the other hand, monetary schemes (penalty and discount) increases the willingness to board the next vehicle.

ABSTRACT. Background: The current non-connected autonomous vehicle scheme for speed changing along the road has limitations. Other alternatives require a central control or a complex communication system between the vehicles. Objective: We suggest controlling the traffic of a line of autonomous vehicles by using predetermined speed profiles. Methods: We introduce a novel method to control autonomous vehicles traffic. Particularly, we investigate cases where specific velocities are required at some points along the road. As traffic flow and speed limit may change due to upcoming road conditions, it is imperative to control vehicle line traffic, such that phantom jams will be prevented while preserving maximal traffic flow at minimum energy. Results: We provide a comparison of these profiles for acceleration, deceleration, and for the combined case in terms of traffic flow, energy consumption, travel duration and the resulting jam characteristics. Conclusions: Following the comparison, we conclude that the best strategy would be to use linear speed profiles both to accelerate and to decelerate. Lastly, we suggest a tool to compare speed profiles for deceleration cases where the formation of an upstream propagated traffic congestion is inevitable.

ABSTRACT. Background The “Risk Compensation” theory (Wilde,1982) argues that every driver has an individual optimal level of risk. The driver’s goal is to keep this level of risk balanced while driving and take action in order to achieve it. For example, if the driver feels that the risk is less than the desired level of risk, s/he would return to his/her desired level by engaging in riskier behavior, usually by increasing speed (Plazman, 1975). Over the recent decade, safety systems installed in cars have become an integral part of them. The purpose of these systems is to decrease the probability of been involved in an accident and/or to minimize the accident’s severity. Considering the increased prevalence of safety systems, it is critical to examine their effect on driving-behavior.

Method We use traffic violation data obtained from the Police and car model (include all characteristics) data obtained from the Ministry of Transportation for the year 2018. Specifically, we use an inflated zero negative binomial regression model, with the existence of Break-Assistance System (BAS) (vs. lack thereof) as our independent variable and speeding tickets as a proxy of reckless driving – our dependent variable. We control for a large number of variables.

Results Our results indicate that BAS is significantly associated with the number of speeding violations across various car models.

Discussion Our results suggest that active safety systems might encourage drivers to engage in reckless driving, as measured by speeding tickets. Presumably, it may lead to more accidents and growing costs. Given the prevalence of car safety systems, drivers and policy makers should be aware of potential hazardous consequences of installing similar safety systems.

ABSTRACT. Driver's performance is crucial for road safety. There is a relationship between performance and stress so that too high or too low stress levels (usually characterized by stressful or careless driving respectively) impair driving quality. Therefore, monitoring drivers' stress levels can improve overall performance by either alerting or intervention during sub-optimal stress levels. Commonly used stress measures suffer from disadvantages, such as delays in indication and sensors' invasiveness. Grip force is a relatively new measure of stress that showed promise in measuring stress during psychomotor tasks. In driving, grip force is a noninvasive, natural, and intuitive sensor as drivers must continuously grip the steering wheel. The aim of the current research is to examine whether grip force can be used as a useful measure of stress in driving tasks. Thirty-three participants took part in two driving experiments. In the first experiment (N=12), participants faced several driving challenges, including slalom driving and unexpected traffic events that required steering and emergency braking. Specifically, emergency braking was found to be the most effective stressor, as braking intensity's effects on grip force stress indices (average, mean, and standard deviation) were found to be statistically significant. In the second experiment (N=21), braking intensity's effects on grip force stress indices were validated. Also, braking intensity's effects on heart rate variability and grip force correlations with heart rate variability were significant. The current research provides initial evidence that grip force can be used as a valid and reliable measure of stress in driving tasks. These findings may have several applications in the field of stress and driving research, as well as in the car safety domain.