Improved Adaptive Cruise Control for Autonomous Vehicles with Consideration of Crash Avoidance

Current Adaptive Cruise Control (ACC) systems are prone to risk of crash from surrounding unexpected cut-in vehicles. Hence, accurate risk evaluation for collisions and corresponding crash avoidance algorithms are highly desired. Therefore, in this work, we propose an accurate Time to Collision (TTC) calculation method using elliptical vehicle geometry, and evaluate the collision risks with surrounding vehicle, quantitatively. To avoid collision with multi-direction cut-in vehicle, a controller-switching mechanism based on TTC is first designed to switch back and forth between the performance-oriented controller and safety-oriented controller. Moreover, the proposed work is validated through simulations. The simulation results reveal that, for different scenarios (car-following, front-side and rear-side cut-in), proposed method can enhance the tracking performance while avoiding crash with surrounding vehicles from front-side and rear-side cut-in, effectively.