Design and experiential test of a model predictive path following control with adaptive preview for autonomous buses. (August 2021)

• Record Type: Journal Article
• Authors: He, Hongwen; Shi, Man; Li, Jianwei; Cao, Jianfei; Han, Mo
• Is Part Of: Mechanical systems and signal processing. Volume 157(2021)
• Publication Date: 2021-08
• Subjects: Autonomous bus -- Lateral control -- Active safety control -- Model predictive control -- Adaptive preview -- Experiment verification; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2021.107701
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: A new PSO-LS-SVM method is developed to mitigate the huge computation burden of the online controller. The feasibility of the PSO-LS-SVM is tested and verified in a real autonomous bus. A hierarchical control system framework is proposed that forces the vehicle to track a predefined path. An active safety controller is designed to prevent the vehicle from sideslip and rollover. Abstract: This paper presents a hierarchical path following control framework for a two-axle autonomous bus, under which the active safety controller, the adaptive preview regulator as well as the predictive trajectory tracker are integrated as a whole system with highly effective performance. The framework is developed with two layers. The upper layer is designed to prevent the vehicle from sideslip and rollover by restricting the speed with an active safety controller. As the high computational load of optimal algorithm poses a great challenge for the real-time application in vehicle system, an adaptive preview regulator is, therefore, developed with an active combination of the PSO and SVM aiming for a more practical implementation. Back to the lower layer, the predictive trajectory controller and PI controller are proposed to acquire the steering angle with stability constrains to follow the reference path. The whole system with the proposed method is verified with both the simulation and experiments. A combined comparison simulation with four cases was carried out based on TruckMaker-Simulink joint platform using a highly-fidelity and full-car model in different scenarios. Compared with the case without preview and that with fixed preview, the tracking accuracy of the proposed control is improved by 44.97% and 36.12% respectively. The proposed method is tested in a real autonomous bus with acceptable errors compared with the simulation. In addition, the effectiveness of the proposed layered control framework and its real-time control capability is proved in real autonomous bus with satisfying performance under parameter uncertainties and external disturbances.