L'invention concerne un procede pour fournir une representation de prediction d'objet, consistant a simuler le deplacement d'un vehicule hote et d'un objet dans un systeme de coordonnees longitudinales-laterales base sur leurs positions initiales au moment du demarrage (t0), un deplacement du vehicule hote et un deplacement de l'objet determine. Des points dans le temps (t0 -n) sont detectes lorsqu'au moins un point de reference du vehicule hote et un point de reference d'objet d'une pluralite de points de reference d'objet ont la meme position longitudinale. Pour chaque point detecte dans le temps, une position laterale associee (y1-n) du point de reference d'objet au point detecte dans le temps est detectee.

Auto-brake systems have been on the market for a few years, but in order to continue improving their performance, multi-target threat assessment is a key component. The goal is to trigger braking as soon as possible, but only if necessary, by detecting when a collision can no longer be avoided via steering. This is done by looking at all objects in the road scene, including other vehicles and barriers, and then looking for escape paths. The paper presents an extension of previous work in this area where escape paths are found in a scenario with generally positioned objects. The escape paths are parametrized in a smooth way that is suitable for both threat assessment, i.e., finding the driver's options in a given situation, and for planning and controlling the vehicle laterally.

There is a strong trend for increasingly sophisticated Advanced Driver Assistance Systems (ADAS) such as Autonomous Emergency Braking (AEB) systems, Lane Keeping Aid (LKA) systems, and indeed autonomous driving. This trend generates a need for online maneuver generation, for which numerous approaches can be found in the large body of work related to path planning and obstacle avoidance. In order to ease the challenge of choosing a method, this paper reports quantitative and qualitative insights about three different path planning methods: a state lattice planner, predictive constraint-based planning, and spline-based search tree. Each method is described, implemented and compared on two specific traffic situations. The paper will not provide a final answer about which method is best. This depends on several factors such as computational constraints and the formulation of maneuver optimality that is appropriate for a given assistance or safety function. Instead, the conclusions will highlight qualitative merits and drawbacks for each method, in order to provide guidance for choosing a method for a specific application.

Auto-brake systems have been on the market for a few years, but in order to continue improving their performance, multi-target threat assessment is a key component. The goal is to early be able to say that a collision can no longer be avoided by a steering maneuver, and thus braking can be started earlier. This is done by looking at all objects in the road scene, including other vehicles and barriers, and then looking for escape paths. The paper presents a computationally efficient way of determining whether an escape path can be found in a scenario with generally positioned objects, given the limited dynamics of the vehicle.

Accident data show that the vast majority of pedestrian accidents involve a passenger car. A refined method for estimating the potential effectiveness of a technology designed to support the car driver in mitigating or avoiding pedestrian accidents is presented. The basis of the benefit prediction method consists of accident scenario information for pedestrian-passenger car accidents from GIDAS, including vehicle and pedestrian velocities. These real world pedestrian accidents were first reconstructed and the system effectiveness was determined by comparing injury outcome with and without the functionality enabled for each accident. The predictions from Volvo Cars' general benefit estimation model are refined by including the actual system algorithm and sensing models for a relevant car in the simulation environment. The feasibility of the method is proven by a case study on an authentic technology; the auto brake functionality in collision warning with full auto brake and pedestrian detection (CWAB-PD). Assuming the system is adopted by all vehicles, the case study indicates a reduction of 24 percent in pedestrian fatalities for crashes where the pedestrians were struck by the front of a passenger car. (A) Paper to the session Driver Assistant Strategies and Accident Causation of the 4th International Conference on ESAR Expert Symposium on Accident Research, 16th to 18th September 2010 in Hannover. For the covering abstract of the conference, see ITRD D366702.

More and more vehicles are being equipped with Automatic Emergency Braking (AEB) systems. These systems intend to help the driver avoid or mitigate accidents by automatically applying the brakes prior to an accident. Initially only rear-end collision were addressed but over time more accident types are incorporated and brakes are applied earlier and stronger, in order to increase the velocity reduction before the accident occurs. This paper describes one of the latest AEB systems called Collision Warning with Full Auto Brake and Pedestrian Detection (CWAB-PD). It helps the driver with avoiding both rear-end and pedestrian accidents by providing a warning and, if necessary, automatic braking using full braking power. A limited set of accident scenarios is selected to illustrate the theoretical and practical performance of this system. It is shown that the CWAB-PD system can avoid accidents up to 35 km/h and can mitigate accidents achieving an impact speed reduction of 35 km/h. To the best of the authors knowledge CWAB-PD is the only system on the market that automatically can avoid accidents with pedestrians.

This paper presents a threat-assessment algorithm for general road scenes. A road scene consists of a number of objects that are known, and the threat level of the scene is based on their current positions and velocities. The future driver inputs of the surrounding objects are unknown and are modeled as random variables. In order to capture realistic driver behavior, a dynamic driver model is implemented as a probabilistic prior, which computes the likelihood of a potential maneuver. A distribution of possible future scenarios can then be approximated using a Monte Carlo sampling. Based on this distribution, different threat measures can be computed, e.g., probability of collision or time to collision. Since the algorithm is based on the Monte Carlo sampling, it is computationally demanding, and several techniques are presented to increase performance without increasing computational load. The algorithm is intended both for online safety applications in a vehicle and for offline data analysis.

This paper presents a new automotive safety function called Emergency Lane Assist (ELA). ELA combines conventional lane guidance systems with a threat assessment module that tries to activate the lane guidance interventions according to the actual risk level of lane departure. The goal is to only prevent dangerous lane departure maneuvers. The ELA safety function is based on a statistical method that evaluates a list of safety concepts and tries to maximize the impact on accident statistics while minimizing development and hardware component costs. ELA runs in a demonstrator and successfully intervenes during lane changes that are likely to result in a collision and is also able to take control of the vehicle and return it to a safe position in the original lane. It has also been tested on 2000 km of roads in traffic without giving any false interventions

Detection and tracking of other vehicles and estimation of lane geometry will be required for many intelligent driver assistance systems in the future. By combining the processing of these two features into a single filter, better utilisation of the available information can be achieved. For instance, it is demonstrated that it is possible to improve the road shape estimate by including information about the lateral movement of leading vehicles.Statistical evaluation is done by comparing the estimated parameters to true values in varying road and weather conditions. The performance is also related to typical requirements of active safety applications such as adaptive cruise control and a new safety function called emergency lane assist.

This thesis is concerned with automotive active safety, and a central theme is a new safety function called Emergency Lane Assist (ELA). Automotive safety is often categorised into passive and acti ...