Coordinated lane change in autonomous driving: a computationally aware solution

Abstract

Abstract This paper addresses the design of coordinated maneuvers in an autonomous driving set-up involving multiple vehicles. In particular, we consider a lane change problem where a vehicle has to merge in a platoon traveling in the adjacent lane of a two-lane one way road. We propose a cooperative solution that trades optimality for computational feasibility without simplifying the merging vehicle dynamics. The key idea is decoupling the problem into two phases: an online coordination phase where vehicles in the platoon create a gap where the merging vehicle can safely enter, and a merging phase, where the merging vehicle change lane by tracking a pre-computed optimal maneuver. A numerical case study shows the achieved trade off between performance degradation and reduction in computing time of the proposed solution.