Sampling-Based Threat Assessment Algorithms for Intersection Collisions Involving Errant Drivers

摘要: Abstract This paper considers the decision-making problem for a vehicle crossing road intersection in presence of other, potentially errant, drivers. is considered game-theoretic framework, where errant drivers are assumed to be capable causing intentional collisions. Our approach simulate possible behaviors using RRT-Reach, modified application rapidly-exploring random trees. A novelty RRT-Reach use dual exploration-pursuit mode, which allows efficient approximation reachability set some fixed time horizon. Through simulation and experimental results with small autonomous vehicle, we demonstrate that this threat assessment algorithm can used real-time minimize risk collision.