After the measurement campaigns have collected data representative of the most common intersection scenarios, the evaluation of the cooperative perception will take place. Then, the developed object-oriented Bayesian networks (OOBN) models will be parameterized and their performance evaluated to account on the quality of recognition of maneuver intentions, situation interpretation and on the overall assessment of risk for traffic situations at intersections. The aim is not to limit the system to solutions for independently modeled single problematic situations, but to use a generic modeling framework, which incorporates changes in all possible states of dynamic objects (traffic participants and traffic lights) in the context of traffic rules, intersection topology and topography. A generic modeling framework allows utilization of the same signal pre-processing module for the same sensor type and treatment of uncertainties in the same framework, with same accuracy. This avoids later on the need of conflict management between "island-solutions" of independently modeled situations in similar context. The advantages of the proposed framework incorporate the proper handling of uncertainties in measurements and situation complexity, efficient combination of various (existing and under development) methods for situation analysis as input to achieve an overall hazard assessment. It is flexible to further extensions of the OOBN-approach to cover more traffic situations, both at intersections (such as obstacles on the entrance- and/or disappear-lanes) and on highway scenarios. Thus, Proactive coaching Safety Assistance System (PaSAS) contributes in general to the road safety, as well as to the project goal on alerting networking road users in critical hazard situations just in time.
Overall probabilistic framework for modeling and analysis of intersection situations
2012
12 Seiten, 4 Bilder, 13 Quellen
Aufsatz (Konferenz)
Englisch
Probabilistic Approach for Modeling and Identifying Driving Situations
British Library Conference Proceedings | 2008
|JOINT PROBABILISTIC MODELING AND INFERENCE OF INTERSECTION STRUCTURE
Europäisches Patentamt | 2016
|