Predictive Force-Centric Emergency Collision Avoidance

Abstract A controller for critical vehicle maneuvering is proposed that avoids obstacles and keeps the on road while achieving heavy braking. It operates at limit of friction structured in two main steps: a motion-planning step based receding-horizon planning to obtain acceleration-vector references, low-level following these acceleration references transforming them into actuator commands. The evaluated number challenging scenarios results well behaved with respect to, e.g., steering angle, body slip, path. also demonstrated successfully balances braking avoidance such it really takes advantage possibilities. Specifically, moving obstacle, makes use widening gap perform more braking, which clear online replanning capability if obstacle should be human or animal. Finally, real-time capabilities are demonstrated. In conclusion, performs well, both from functional perspective perspective.