Maneuver-based risk assessment for high-speed automotive scenarios

This work presents a novel approach for collision assessment for automotive environments. As collision prevention and risk analysis are key challenges for today´s intelligent transport systems, sophisticated solutions for a collision-free trajectory generation get indispensable. The presented collision checker is integrated into an optimal control based planning framework that generates minimum jerk trajectories for arbitrary maneuvers. In contrast to common collision checkers, the developed method does not need to discretize the time space, but gives an algebraic solution which covers nearly all situations and has a constant response time. Due to its fast evaluation it is predestined for use in a Monte Carlo simulation respecting the probabilistic nature of real world traffic scenes. An example implementation of the proposed method is applied to simulation scenarios that demonstrates the benefits to comparable approaches. The obtained results have proven the feasibility of our approach to risk analysis of traffic situations.