Coordination of Automotive Active Safety Systems via Global Chassis Envelope Control

Abstract

Ground vehicles, characterized by their inherently over-actuated nature, serve as a foundation for the development of hierarchical control frameworks and control allocation schemes, which may be strategically harnessed and developed to achieve effective handling and enhanced traction on road surfaces via the coordination of their multiple active safety systems including active front steering (AFS), differential braking, and active suspension (AS).

To adhere to global chassis control, this thesis first introduces a global fuzzified model predictive control allocation (MPCA) scheme, integrating AFS, differential braking, and AS whilst considering their actuator dynamics. Fuzzy logic membership functions are set up to facilitate the appropriate coordination between these active safety systems.

Thereafter, stability regions defined in phase portraits connecting side slip and yaw rate are integrated into a model predictive controller and are used to derive a dynamically-changing allocation index. The decision to employ phase portraits stems from their ability to visualize ground vehicle dynamics and stability, which will be first used to enhance the lateral stability of ground vehicles within a hierarchical control framework. This results in effective resource allocation between front steering and braking, and thus significantly enhances lateral stability.

This research extends its focus onto explicitly defining and developing three-dimensional stability regions in the diagram relating the lateral and longitudinal accelerations to the yaw rate of the vehicle. This allows for the coordination of the active safety systems through a novel control architecture that employs this stability envelope in the design of the upper controller and within the control allocation scheme to attain the desired notion of global stability. This advancement not only addresses rollover risks but also improves both lateral and longitudinal handling performance in ground vehicles, whilst analyzing and employing measurements directly taken from the inertial measurement unit, thereby eliminating the need for state estimators in the stability regions.

To evaluate the proposed approaches, extensive simulations are conducted in the CarSim-Simulink environment for the various control schemes in corresponding appropriate testing maneuvers to verify the holistic design involving the a prediction of the actuator dynamics and involving a stability envelope on the global level of the vehicle. First, the obtained results demonstrated the advantages of integrating the actuator dynamics into the allocation scheme, which involve an anticipation of the effect of the actuator's dynamics on the control input. Specifically, 41.89 % and 36.38 % lower errors along the yaw rate and side-slip angle are the results of designing an MPCA scheme involved with a fuzzy logic system for a double lane change maneuver. The consideration of a 2D safety envelope in a model predictive control scheme demonstrates superior stability and handling performance, resulting in 20.6 % and 65.7 % smaller sideslip and yaw rate tracking errors, respectively, over the bench-marked nonlinear model predictive controller. The natural progression onto 3D to achieve global stability bestows a 62.63 % decrease in slip and 69.18 % lower roll angle in comparison to the commercially available controller.