Path-tracking and lateral stabilisation for autonomous vehicles by using the steering angle envelope

In the design of a path-tracking controller for autonomous vehicles, mediating the conflicting objectives among path-tracking, vehicle stabilisation, and low computational cost is a challenging issue. Accordingly, this paper proposes a model predictive control (MPC)-based path-following controller with steering angle envelopes. In contrast to previous MPC-based structure, in which the constraints in terms of the road sides and lateral stabilisation directly impose on the algorithm, in the proposed approach, these aspects are formulated as the steering angle envelopes. This feature enables the controller to simultaneously ensure the tracking performance and computational feasibility. Moreover, considering the trade-off between path-tracking and vehicle stabilisation, the full-length prediction horizon pertaining to the steering angle envelopes are divided into two different parts: the safe road envelope is applied in the short-term, and the stable handling envelope is applied in other part. By modelling vehicles driving under different road conditions with various driving speeds and road adhesion values, the developed controller is compared with several existing control schemes via simulations and hardware-in-the-loop (HIL) experiments. The comparison results demonstrate the effectiveness of the proposed controller in realising path-tracking, lateral stabilisation, and real-time computational capabilities in the best possible manner.

Automated summary

This paper proposes a model predictive control-based path-following controller with steering angle envelopes, balancing path-tracking, vehicle stabilisation, and low computational cost. The controller divides the full-length prediction horizon into safe road envelopes for short-term application and stable handling envelopes for longer durations. Through simulations and hardware-in-the-loop experiments, the controller is shown to effectively achieve path-tracking, lateral stabilisation, and real-time computational capabilities under various driving conditions.