LAUSR

Abstract Solving the problem of intersection crossing for autonomous vehicles is a challenging task due to combined combinatoric and dynamical control decisions. To reduce the complexity of the computations and distribute the resulting global optimization problem, we propose a combined scheduling-control method. Thereby, in this paper, we focus on the formulation of a resource-constrained-project-scheduling problem (RCPSP) to solve the combinatoric decision, i.e. the order in which vehicles cross an intersection area in a central coordination unit. This problem considers control decisions from the vehicles, which are computed using model predictive control (MPC) laws. In turn, the resulting scheduling solution is incorporated again in local vehicle MPC problems, which negotiate among each other to find a dynamically feasible solution. This seamless combination of scheduling and control results in efficient solutions, which is illustrated using numerical simulation and the results are compared with a first-come-first-served (FCFS) strategy.