Vertical trajectory planning: an optimal control approach for active suspension systems in autonomous vehicles

Abstract

In this paper, a new control algorithm for active suspension systems is presented. Latter is designed to improve low frequency comfort, which gains importance with autonomous driving and the possibility for nondriving tasks like in-car working. The algorithm is based on model predictive controllers and makes use of environment sensor information and extends preview by taking the future trajectory of the vehicle into account. A simulative benchmark demonstrates the advantages compared to existing control algorithms. Vertical- as well as roll- and pitch-acceleration can be significantly reduced by at least 20% compared to a non-preview controller up to a frequency of 5 Hz. A prototype vehicle implementation of the new controller enables a subjective and objective review of the resulting vehicle motion and confirms the results of the simulation. Despite of limitations on the precision of the preview information, vertical acceleration could be reduced by 5.5% in this implementation. It was shown that a future series implementation is possible and can be used to improve autonomous driving comfort by increasing low frequency suspension performance and efficient integration of additional features such as curve tilting.