Optimisation of high-speed crash avoidance in autonomous vehicles

Abstract

In the context of a future scenario of autonomous vehicle platooning, this paper considers the optimisation of a vehicle’s standard brake, acceleration and steering control inputs, for collision avoidance. We consider the case where escape into the neighbouring lane is feasible. An iterative simulation-based method is used, which allows vehicle parameters to be optimised simultaneously; this also allows us to find the best vehicle handling balance for such a manoeuvre and to quantify the cost of suboptimal design. The paper also considers the relative advantages of speed reduction in conjunction with rapid lane change. The goal here is to quantify the best possible vehicle escape manoeuvre and the relative cost of alternative strategies. The paper does not provide an immediately practicable controller, but simple open-loop approximation of the optimal controls suggests a route towards future real-time solution.