ABSTRACT. With the number of in-vehicle information systems and the complexity of their tasks growing at a very high rate in near future, we need a clear understanding of their related distraction or mental workload and its impact on driver performance as these systems do not always comply with the intended driver safety enhancement. The advent of vehicle automation promotes a reduction of the driver workload. However, depending on the automation grade; consequences for the passengers such as out-of-the-loop states can be foreseen. This presentation gives an overview of the impact of such technologies on traffic awareness for the driver towards improving driving performance and reducing road accidents. Furthermore, the benefits and potential problems regarding vehicle automation will be outlined.

ABSTRACT. Spacing policy for autonomous vehicles is one of the important issues because it is highly related to the safety on the roads, efficiency of vehicles, and user satisfaction of autonomous vehicles. Many researchers have developed several spacing policies for autonomous vehicle control. Most of the spacing policies are mainly focused on the control scheme such as string stability and safety. So, the issues related to user satisfaction and possible effect of autonomous vehicles on manual vehicles are less considered especially in the mixed traffic situation of autonomous and manual vehicles. In this study, we propose an Asymmetric Collision Risk(ACR)-based spacing policy based on the analysis on collision risk and driving behavior. The proposed spacing policy is compared with other spacing policies by simulating with trajectories of human drivers. Based on the simulation results, the performance of each spacing policy is compared in terms of vehicle operation, CO2 emission and safety. The results show that ACR spacing policy has similar pattern with the human driver with smoother trajectory and less acceleration/deceleration actions. In terms of road efficiency and environment, the proposed ACR spacing policy shows the second best performance next to Safety Spacing policy. However, in terms of safety in mixed situation of manual and autonomous vehicles, ACR spacing policy is superior to other spacing policies with zero occurrence of critical event. By considering overall performance, the proposed ACR spacing policy is expected to generally show good performance in the mixed traffic situation of manual and autonomous vehicles.

ABSTRACT. In this paper, we consider the coordination problem of multiple autonomous vehicles at traffic intersections. In particular, we exploit a cooperative, sequential conflict resolution approach based on a pre-defined decision order. Using an optimal control formulation, we show how coordination can be ensured by solving two local problems where collision avoidance is enforced as time-dependent state constraints. We will analyse the feasibility of a given sequence with respect to different decision criteria and present simulation results supporting our results.

ABSTRACT. Lane change is a complicated maneuver and a reason for many severe highway accidents. Automatic lane change has great potentials to reduce the number of accidents. Previous researches mostly tried to find an optimal trajectory that can be applied for simple lane change. They do not consider time/distance constraints for doing the lane change during the merging/exiting. Through analysis of human driver lane change data, we propose a multi segments behavior and motion model to mimic the human driver operation. We developed a simulation platform in PreScan and evaluated proposed automatic lane change model for challenging scenario in the merging/exiting.

ABSTRACT. Cameras are an essential component of many Advanced Driver Assistance Systems (ADAS) such as Lane Departure Warning, Lane Keeping Assistance, Automatic Emergency Braking, Traffic Sign Recognition and Night Vision. With the ongoing evolution of ADAS towards Automated Driving, the amount of cameras within vehicles as well as the performance of these cameras is only expected to grow. Nowadays automotive cameras are typically sensitive beyond the visual range (i.e. into the near infrared range) They may also use up to 20 bits precision for their internal signal. In order to support the development process of ADAS and Automated Driving Systems by means of simulation, it is required to have detailed physically correct simulation models of these automotive cameras. This paper describes the implementation and validation of a new spectral camera sensor model in the ADAS simulation platform PreScan. The traditional RGB-based camera implementation created images limited to the visual spectrum. In contrast, the new PreScan spectral camera sensor model allows for high color depth, broad spectrum and physically correct image generation. The model provides a real-time simulation of light in the 320-1500nm spectral range which traverses through the various media and lenses before the sensor produces the final image signal. Other important features of this new detailed camera model are a depth resolution of 20 bit per waveband as well as an increased dynamic range.

ABSTRACT. This paper proposes a method that uses a state-space model to predict vehicle behavior. It incorporates a proposed model of a generic driver, who is described by a simple feedback model with various constraints corresponding to general traffic rules and the physical limits of the vehicle. In addition, data assimilation is used to estimate the state and the model parameters on-line. By being able to predict natural driving data in unknown situations, the prediction error of the proposed model is signicantly reduced compared to constant prediction, since data assimilation can be used to adapt the model to new situations. The data assimilation makes this model universal and robust against any new road-traffic situations. Furthermore, itis confirmed that this ability to predict can improve the fuel efficiency of adaptive cruise control systems.

ABSTRACT. This paper presents 2-D pedestrian motion prediction model based on potential field for a hazard-anticipatory collision avoidance braking system to enhance the collision avoidance performance. 1-D pedestrian motion prediction reaches its limit under the situation that a pedestrian walks towards a parked vehicle as the prediction is conducted based on the current position and velocity of pedestrian within a finite time horizon. If the change of pedestrian motion can be predicted, the vehicle maneuver to avoid collision with pedestrian can be executed in advance. 2-D pedestrian motion prediction model based on potential field is constructed. The effectiveness of the proposed prediction model is verified by comparing the measured trajectory with the calculated data based on 1-D prediction and 2-D prediction.

ABSTRACT. Recently, demand for active safety systems has been increasing toward zero traffic accident fatality, thus an automated emergency brake system to avoid pedestrian accidents will be commercialized corresponding to Euro-NCAP. However, an effective assistance system for pedestrian accidents while turning at an intersection has not been proposed, though number of the accidents is the second largest after crossing pedestrian accident. We propose the driving assistance system utilizing the surround monitor camera system which is getting popular. Stereo camera configuration utilizing front and side fish eye cameras has been considered to improve detection reliability. However, particular concerns for this configuration are existed, thus we are trying to develop the solution applying machine-learning feature amount in this research. They were preliminarily verified utilizing computer graphics, and then experimented for actual images captured in the real world.

ABSTRACT. The braking behaviors of drivers when a pedestrian steps out from a sidewalk and into the street were analyzed using a driving simulator. Based on drivers’ braking behavior, a braking control timing for a system for avoiding collisions with pedestrians was proposed. In this study, the subject drivers started braking at almost the same time in terms of time to collision (TTC), regardless of the velocity of the subject vehicle and the crossing velocity of the pedestrian. The results of our experiments showed that, to minimize the degree of interference between the driver and the system, the optimum timing at which to apply the braking was at a TTC of 1.3 s. Next, the drivers' braking behavior wa  investigated when the system controlled the braking to avoid a collision with this timing. The drivers did not exhibit any change in their braking behavior, while there was no excessive interference between the braking control system and the drivers when attempting to avoid a collision with a pedestrian, indicating that the drivers were not becoming excessively dependent on the system.

ABSTRACT. This paper proposes fast lane visibility check methods for traffic prediction using traffic lane information and measurement information given by range sensor. Especially, this paper describes high accurate positioning method using high density point cloud map, competing lane search method, and fast visibility check method using high precision vector map and 3D LiDAR(Light Detection and Ranging). Finally, this paper illustrates experimental results in urban environments. In this experiment, the proposed method can detect occluded traffic lanes at crossing points, behind of vehicle and slopes.

ABSTRACT. In this paper we describe an approach which combines data from lidar sensors and acoustic sensory data for an improved object tracking accuracy in the setting of an autonomous car. After explaining the sensory setup we chose, we will show how acoustic data from two Kinect cameras, i.e., multiple microphones, which were mounted on top of a car, can be combined to derive an object's direction and distance. Part of this work will focus on a method to handle non-synchronized sensory data between the multiple acoustic sensors and the lidar sensor respectively. In a second step we will present a method which extends the given approach in order to identify and to localize cars with a certain sound signature, e.g., emergency vehicles. We will give experimental results for real traffic scenarios.

ABSTRACT. The personal mobility vehicle (PMV) described in this paper has typical biomimetic characteristics, which incline the vehicle body inside a curve. Since the behavior of this vehicle is similar to that of a two-wheeled vehicle, it is necessary to examine the PMV and the potential for interference with other vehicles in the midst of ordinary road traffic. Therefore, we simulated vehicle dynamics (including a driver model and a road-transportation system) and then analyzed the results, to guarantee the PMV’s compatibility with micro-scale traffic flow.

ABSTRACT. Information about the vehicle environment becomes increasingly important as the development of automated driving functions progresses. This paper not only addresses a perception of the current environment, but also an estimation of possible future states. A multi-purpose method for measuring objects, lane markings and road boundaries using a reference sensor system that includes laserscanners is presented. The focus lies on recording a comprehensive environment description that is needed for the examination of vehicle interaction. The results are utilized in the correspondent tool chain for postprocessing, event detection and analysis for object prediction applications. In a final step, different possible applications are introduced.

ABSTRACT. Conventional driver assistance systems suffer from relatively late reaction timing. Adaptive Cruise Control (ACC) systems, for example, react to vehicles cutting-in from neighboring lanes once these vehicles are at least partly driving on the host vehicle's lane. The improvement by state of the art prediction approaches is limited, because the reliable approaches base on features, which are characteristic for the phenomenological effect of a behavior change. That means they are capable of predicting a behavior once it started. The Honda i-ACC overcomes the above mentioned limitation of delayed reaction by using a novel approach of behavior prediction. This approach combines two prediction methods: A context based prediction (CBP) and a physical prediction (PP). The CBP evaluates the situational context of a predicted vehicle and does not use any phenomenological feature. It is thus capable of long-term prediction. In contrast, the PP evaluates phenomenological features by accumulating the predicted vehicle's history of recent positions and comparing them to a set of trajectories, which allows for an accurate short-term prediction. In an evaluation on 15000km of driving in Europe we will show that the new prediction approach achieves both, reliable and long-term prediction.

ABSTRACT. Separably developed functionality as well as increasing situation complexity poses problems for building, testing, and validating future ADAS. These will have to deal with situations in which several current ADAS domains interplay. We argue that a generalized estimation of the future ADAS function benefit is required for efficient testing and evaluations, and propose a quantification based on an estimation of the predicted risk. We show that such an approach can be applied to several different types of risks and to such diverse scenarios as longitudinal driving, intersection crossing and lane changes with several traffic participants. Resulting trajectories exhibit a proactive, ''foresighted'' driver behavior which smoothly avoids potential future risks.

ABSTRACT. This study examines the driver assistance system which informs the passage possibility at a signalized intersection ahead. The assistance system indicates the distance which the vehicle can forward by maintaining the present velocity until turning to the red signal using the signal information. If the indicated distance is shorter than the distance to the intersection from the vehicle, the system also indicates a stopping distance by assuming the ordinary deceleration. The driving simulator experiments are conducted to evaluate the proposed assistance system. The assistance system encourages the earlier deceleration before the amber signal and prevents the emergency braking behavior or the risky passage through the intersection.

ABSTRACT. This paper evaluates the effectiveness of an active chassis control function after an initial light impact. The experiment was designed in a high-fidelity driving simulator with motion platform that is adapted to mimic a real-world crash. The post-impact control function here superimposes individual-wheel brake and steering wheel torque in order to minimize the vehicle maximum lateral deviation from its original traveling lane. It is found out that drivers with the function intervention has smaller lateral deviation after the disturbance from an initial impact, and thus less risk of leaving the road which may cause a secondary impact. The driver reaction with and without the function is also found to be different in terms of both braking the steering behaviours.

ABSTRACT. The aim of this paper was to extend an existing simulation environment for lane departure accidents in the USA and Japan to be applicable for Germany. The German situation was described primarily using the German in-depth accident Study (GIDAS) as input to the simulation. Simulation output was validated against accident data. Two changes to the simulation environment were proposed: First, use of weighted empirical cumulative distributions for input variables, and second, use of weighted logistic regression for injury risk curves. Further methodological consideration, should be given to inclusion of non-injury cases, object grouping and treatment of uncertainty.

ABSTRACT. Car manufacturers and automotive suppliers design more and more Advanced Driving Assistance Systems (ADAS) that are quickly installed into cars. Nevertheless, such ADAS address different types of driving activities and driver’s behaviors; they might concern longitudinal and lateral controls, lane changes, navigation, and try to take into account driver state and behavior. But what about the complementarity of such different ADAS if they are all together implemented in one car? And what about their interactions with the driver? Interactions have to be defined according to the levels of automation selected by the car manufacturer. In the French national project CoCoVeA (French acronym for Cooperation between Driver and Automated Vehicle), three levels of automation have been studied involving several existing ADAS. And this paper proposes a methodological approach to identify competences and capacities of driver and ADAS in the case of many driving situations in order to check complementarity, function allocation, authority management, as well as cooperative aspects in order to assess reliability of such a Human-machine system and the acceptance and even the attractiveness of highly automated vehicle.

ABSTRACT. Autonomous driving vehicles will become a key element of mobility in the near future. Although it opens up new opportunities for the individual and increases traffic safety, there is a highly controversial debate about how much this new way of driving is accepted by the driver. In this article, relevant factors and expectations for autonomous driving are investigated. The basic theories on determinants for driving enjoyment will be opposed to different levels of automation for vehicles to predict their possible effect on driving enjoyment. An insight will be given into content analysis and online survey. Results indicate that attitudes towards autonomous driving can vary between situations. Especially situations with a high dynamic aspect such as driving a curved road are rejected, while over 90% of the participants were able to name at least one situation where they could imagine to be driven with an autonomous vehicle. This paper contains a detailed presentation of the results and is completed with the discussion of possible impacts of autonomous driving on the driver, on society, and its potential implication for the automotive industry.

ABSTRACT. This paper investigates model complexity, or equivalently, necessary model structure, for describing car-following behavior. The recently suggested PrARX modeling framework is systematically investigated and compared with simpler model structures containing less tuning parameters to see to what extent complex models can be motivated. The results indicate that for long prediction horizons there is no gain having a complex model structure. However, one-step ahead prediction of complex models can be useful for classication of whether the driver should be braking or not.

ABSTRACT. The submitted abstract presents our methodology for designing interface for autonomous driving. The methodology is based on a cognitive approach, the Cognitive Analysis, which permits to extract Augmented Reality information that will convey drivers'attention to particular points of interest in the environment. It will be evaluated on lane changes and handover scenarios.

ABSTRACT. Motorcycle and bicycle Anti-Roll-down System, that controls vehicle roll angle by gyro moment created by the directional changes of flywheel axis, is proposed. Control system is designed by referring the vehicle motion model and effects of gyroscopic torque on vehicle motion. The effects of the proposed system is evaluated and validated by theoretical analysis and vehicle motion simulation.

ABSTRACT. The automotive mobility innovative systems like: Advanced Driver Assistance Systems (ADAS), Active Safety Systems, Autonomous Driving and even Connected Vehicles are continuously being developed for more convenient driving and for achieving higher safety standards. Therefore, software and hardware designs have to handle the increasing challenges and complexities introduced by those systems. Interactive distributed automotive real time systems are typically used to implement such complex systems, which have many real-time constraints distributed in several Electronic Control Units (ECUs) and communication bus(es). This requires form OEMs to specify accurate software integration performance requirements for different suppliers, which is very tough task and contains many integration challenges. This talk will discuss the new trends in the embedded systems software design to overcome the challenges introduced by active safety and autonomous driving systems. The state of the art industrial solutions in software real-time architectural design will be discussed like ATESST2 project and Timm-2-use project. Moreover, the AUTOSAR timing extension standard (used to describe timing specifications, real-time constraints and probabilistic timing properties for events and event chains) will be explored as a good practice to provide clear timing requirements for each ECU. It also provides sufficient timing information for real time measurements and for efficient validation of the system real time requirements. A practical example of BMW tools (Artime, T1 and Artip), as mentioned in BMW’s publications, for AUTOSAR timing extension will be shown. Moreover, some other advanced real-time architectural design methods shall be discussed like: modeling the timing requirements and real-time constraints, modeling the real-time architectural design (operating systems tasks and interrupts), simulating the real-time architectural design, optimizing & refining the real-time architectural design, verifying the timing requirements and time triggered based real-time Architecture. Recent approaches predict the tasks and ISRs execution times used in the real-time architectural designs at early phases of the development. One of the approaches uses machine learning algorithms and provides good estimates of the tasks’ execution times and the overall CPU loads. Afterwards, the complete real-time architectural design simulation and verification is used to be able to refine and optimize the design. Moreover, eVlaue, European project for evaluation of active safety systems, will be explained with its performance testing process and the impact on software design. Active safety and self-driving systems are increasing the usage of sensors fusion (e.g. fusion between laser scanners, radars, cameras, and ultrasonic sensors) and the usage of high speed buses (e.g. Ethernet, MOST) to exchange data between sensors and their control units. Therefore, combination of some embedded hardware systems are needed (e.g. multi-core, FPGA, DSP processors, High speed automotive buses, etc..). This talk will explain the impact of using those complex hardware systems on software design taking Audi zFAS computer, shown in CES 2014 in Las Vegas, as an industrial solution example.

ABSTRACT. This paper deals with a fundamental study to analyze driver characteristics using driver model. At the first stage of this research, we consider a possibility for identification of driver control manner by using a preview driver model. Influence of each term for preview in the model has been considered by using experimental data, and the order of the model has been decided. And it is shown that the model above can describe an ordinary control action of a driver. From these results, it is shown that the driver model constructed in this research is effective for driver’s various analyses.

ABSTRACT. Traffic accidents are commonly found on horizontal curves. It is therefore important to study how the curve geometry affects the driver behaviour. This paper focuses on analysis of speed and maximal lateral acceleration in seven curves on two-lane rural highways in Sweden. The curve geometry factors studied are radii, presence and length of spiral transitions, tangent lengths and radius of previous curve. Of the studied factors, radii and spiral transitions were found to influence the driver behaviour most. Both larger radii and longer spiral transitions result in higher speeds in curves, and speed variations within curves seemed to be independent on choice of speed entering the curve.

ABSTRACT. We investigated factors determining the start position, defined as the point where a driver presses the gas pedal immediately prior to entering an intersection. Start positions are highly important for safety, because they mark the point where drivers complete their safety confirmation and begin accelerating to enter an intersection. Drivers’ intersection crossing behavior, without right-of-way, was recorded at six unsignalized intersections in a residential area of Tsukuba, Japan. A total of 8 subjects drove the vehicle, at each intersection, 36 times. We analysed only the first three and last three trials because our previous study showed that start positions changed over repeated trials. We used linear regression analysis to identify significant explanatory variables. For the last three trials, own-road width (p < 0.01), crossroad sidewalk width (p < 0.01), and driver properties such as the start position alteration
pattern (p < 0.01) were significant. The adjusted R-squared was 0.60, and the root mean square error was 0.80. The distance from the nearest edge of the crossroad to the start position typically increased when the crossroad sidewalk width and/or own-road width were broad, and was roughly equal to or less tha
the crossroad sidewalk width. For the first three trials, however, only own-road width (p < 0.01) was significant, and crossroad sidewalk width and the start position alteration pattern were not significant. The R-squared was 0.48, and the root mean square error was 1.29. These results suggest that subjects didnot grasp the dimensions of the intersections sufficiently, in approaching the crossing, to drive as effectively as they otherwise would. Through repeated trials, they learned these dimensions, and drove more effectively as a result. This suggests that drivers should be better informed of intersection dimensions, for improved safety. The statistical significance of the start position alteration pattern indicated that individual drivers tended to maintain their characteristic driving behavior when crossing the intersections, which clearly demonstrated the importance of driver behavior properties as explanatory variables. It seems important, then, to consider this pronounced tendency when seeking to provide suitable assistance to individual drivers, with the aim of improving driver safety.

ABSTRACT. The Vestibulo-ocular reflex (VOR) model proposed by Merfeld and Zupan (2002) has been used in a wide number of medical applications as well as the driver behavior model. This model is one of the reflex eye movements. It can deal with the interaction between the otolith and the semicircular canal with four parameters to compensate for the individual difference of the VOR characteristics and two parameters to compensate for that of the eye muscle characteristics. In order to increase the reliability, exact ability of this model, by using genetic algorithm (GA) the parameters were identified base on the result of experiment and simulation. We conducted 12 experiments with the motion capture and the eye movement equipment. The new parameters identified by the new GA technique improved simulation results compares with Merfeld parameters and the previous GA method by applying the generating near optimal initial population and changing fitness function. The range of each parameter on VOR model was identified with main purpose to choose the best parameters for each model application. We developed a Matlab toolbox for identifying parameter base on Matlab Graphical User Interfaces.

ABSTRACT. In this paper, a numerical driver behaviour model is presented in terms of how drivers possibly regulate their deceleration in braking tasks. It is based on David N. Lee’s τ-theory, which describes how drivers can perceive values of relative motion like time to collision or headway. This physiological model provides more universal application possibilities than purely empirical models, but its validation within critical traffic situations constitutes great issues. Therefore, a highly dynamic driving simulator providing all the requirements needed is currently under development at the Dresden University of Technology.

ABSTRACT. This study proposes a method to extract the unique driving signatures of individual drivers. We assume that each driver has a unique driving signature that can be represented in a k dimensional principal driving component (PDC) space. We propose a method to extract this signature from sensor data. Furthermore, we suggest that drivers with similar driving signatures can be categorized into driving style classes such as aggressive or careful driving. In our experiments, 122 different drivers have driven the same path on Nagoya city express highway with the same instrumented car. GPS, speed, acceleration, steering wheel position and pedal operations have been recorded. Clustering methods have been used to identify driving signatures.

ABSTRACT. The AstaZero proving ground for testing of vehicle active safety systems was inaugurated in August 2014. Since then, several research projects have already been active on the proving ground. This paper will introduce the test track, give examples of such research activities, and show the benefits of an open facility for active safety research.

ABSTRACT. A failure-sensitive driver model has been developed in the research study presented in this paper. The model is based on measurements of human responses to different failure conditions influencing the vehicle directional stability in a moving-base driving simulator. The measurements were made in a previous experimental study where test subjects were exposed to three sudden failure conditions that required adequate corrective measures to maintain the vehicle control and regain the planned trajectory. A common driver model and a failure-sensitive driver model have been compared, and results for the latter agree well with the measured data. The proposed failure-sensitive driver model is capable of maintaining the vehicle control and regaining the planned trajectory similarly to the way in which humans achieved this during a wheel hub motor failure in one of the rear wheels.

ABSTRACT. We believe that a driver’s operation and vehicle’s motion characteristics during a lane change differ by driver. This study examined a driver's acceptance and operation in relation to a change in the vehicle’s response using a four-wheel steering vehicle. We changed the frequency response characteristics of the yaw rate for the steering wheel operation by changing the control parameters of the four-wheel steering system. In addition, we conducted some experiments using this experimental vehicle. The results show that the vehicle response that was acceptable to the driver was characterized, and the behavior of drivers in relation to the vehicle response was determined.

ABSTRACT. The premise of driving support systems is to reduce a driver’s load and traffic accidents as a whole. However, we feared that such systems can degenerate a driver’s ability by causing an overreliance on driving assistance systems consciously and/or unconsciously. Therefore, we produced a pre-reminder system that reduces encounters with accident factors by helping the driver turn their attention back to safe driving methods. We installed the system in a driving simulator and conducted an experiment to assess the system. As a result, the pre-reminder system did work as intended, and the driver’s behaviour was focused more on the road, lessening overreliance on road safety systems.

ABSTRACT. In Japan, many head-on collisions occur on subsidiary roads at non signalized intersections and elderly drivers tend to cause this type of accidents. Therefore, to clarify the principal accident causation of head-on collisions is strongly required in our country which is an aging society. In the current study a simulation program was established based on the experimental data and the driving behavior of expert drivers and elderly drivers were compared with focusing on the collisions with crossing bicycles, drivers’ confirming safety behavior and their foot positions.

ABSTRACT. Mitsubishi Electric has been developing an automatic parking system. The main part in the system is provisionally named e-Park controller. It executes PSM (parking slot measurement) using our sonar sensors to find a parking slot available. Simultaneously, the controller estimates states of the vehicle and calculates an optimal path to the parking slot. After the path planning, reference signals for our electric power steering and driving motor control systems are generated to track the path automatically. The parking can be executed externally using our device for the keyless entry system. For safety, the parking system also has an emergency braking function based on the sonar sensors and operation of the driver. This paper mainly focuses on the design of the path planning algorithm and the vehicle speed controller. The automatic parking system is verified in experiments.

ABSTRACT. Driver drowsiness is one of major causes of fatal traffic accidents. We proposed a driver assistance system with a dual control theoretic scheme, which tries to perform safety control of a vehicle as well as identification of a driver’s state (e.g., low arousal, degradation of mental and physical states) simultaneously. The dual control theoretic driver assistance works in a situation-adaptive manner, and it takes into account sleepy driver characteristics and circumstantial contexts. This paper gives an analysis of driver behavior when the driver assistance is active, and discusses the type of assist that is effective for prevention of sleep-related vehicle accidents.

ABSTRACT. This paper describes an approach to vehicle motion control algorithm for application of an automated vehicle. A full recognition of environment is achieved by data fusion and data interpretation based on the dynamic measurements from the environmental sensors such as radar and vision. The recognition of environment is transformed into an artificial risk potential representation, which is collision probability. In order to guarantee lateral stability in the high speed maneuver, constraints of the vehicle motion based on vehicle dynamics have been considered. Probabilistic predictions of surrounding vehicles are included into the risk potential function for represent the predicted collision risk based on the estimated motion of the subject vehicle and the surrounding vehicles. The potential field interpretation has been proposed to represent the risk potential to the surrounding vehicles based on the predicted motion of the vehicles. The proposed motion control algorithm has been investigated via closed-loop simulation.

ABSTRACT. This paper presents a probabilistic threat assessment of vehicle states using wireless communications for application to integrated risk management system. Since one of the main concerns in risk assessment is an assessment of collision risk, it is necessary to estimate the probability of collision occurrence of the ego vehicle which expresses driver sensitivity. Recent developments in vehicle onboard computers and wireless communications devices, also known as dedicated short-range communication (DSRC) devices allow the exchange of information between vehicles (V2V communications). The prediction performance of the proposed method, which integrates V2V communications to conventional advanced driver assistant systems (ADAS), has been investigated via simulations using the commercial vehicle software Carsim and Matlab/Simulink. It is shown the threat assessment performance for the given driving situations can be significantly enhanced by the proposed algorithm. This leads to improvement of conventional driver assistance functions of ADASs.

ABSTRACT. A simulator study was conducted to characterize driver steering responses to lane departure warnings. Drivers were asked to perform a distracting secondary task while their vehicles were occasionally pushed to the side of the roadway to encourage lane departures. Three lane departure warning treatments were investigated: no warning, an auditory warning, and a tactile warning using steering wheel vibration. Overall drivers’ corrective steering responses suggested that both the presence of a lane departure warning and characteristics of the roadway environment influenced the magnitude of the steering corrections observed.