Quantifying performance of a connected vehicle by optimal control

Abstract In this paper we exploit trajectory optimization techniques to quantify acceleration performance improvements in an advanced powertrain that are achievable if a preview of a launch event is available about 1 sec in advance through vehicle to traffic light communications. The vehicle is at idle immediately before acceleration. Based on the preview information, the idle speed is increased immediately before launch. A dynamic optimization approach is employed to optimize the idle speed set-point and trajectories of clutch and engine actuators after start of the acceleration. The optimization is applied to a high complexity model of powertrain. To perform the optimization, the trajectories being optimized are first transcribed using Gaussian kernel functions and then a mesh adaptive direct search (MADS) algorithm is applied. The achievable average acceleration with and without preview (i.e., without raising pre-launch engine speed) and for different sets of actuator trajectories being optimized under drivability constraints are compared thereby quantifying the benefits of vehicle-to-infrastructure connectivity.