IMPLEMENTATIONS OF CAR-DRIVER BEHAVIOUR AND PSYCHOLOGICAL RISK MODELS --ROAD USER BEHAVIOR. THEORY AND RESEARCH. PAPERS PRESENTED AT THE 2ND INTERNATIONAL CONFERENCE ON ROAD SAFETY HELD IN GRONINGEN, NETHERLANDS, AUGUST 1987

摘要

In 1985 a project was started to apply ai techniques to modeling traffic behaviour. One of the first goals in this project was to implement a simulated world in which robots were able to negotiate same-order intersections by taking correct course and speed decisions. The current simulation consisted of three layers. The function of the bottom layer was to control the dynamics of the real world. It computed the coordinates of the objects that were currently moving, the speed, acceleration, the layout of the road, and visual obstructions. It also controlled whose turn it was to look and think. The second layer was the data driven camera. The camera was pointed in a certain direction, computed information, and filled the central database of the robot that was thinking at that time. The third layer consisted of the expert that used the information provided by the camera. The expert decided on his next sight direction, speed and course, based on his intended destination, information on other robots and, (official and psychological/heuristic), traffic rules. A high level algorithm for overtaking was presented. The algorithm was used to demonstrate that in an actual simulation, risk-handling can be a factor on all task levels. It was suggested that simulation techniques introduced a new and fruitful dimension to modelling traffic behaviour.(a) for the covering abstract of the conference see IRRD 815404.