Control of tracking stability for industrial vehicles under low-to-medium adhesion conditions based on LTI/LTV-MPC

Abstract

Unmanned forklifts operate in highly complex environments, and their distinctive suspension systems often lead to lateral load transfer during steering maneuvers. This phenomenon markedly compromises tracking accuracy, particularly on surfaces with low-to-medium adhesion. To mitigate this issue, this study proposes a path tracking controller for unmanned forklifts, leveraging the estimation of transient tire cornering stiffness. The controller comprises three integral modules: a transient cornering stiffness estimation module, a cornering stiffness state prediction module, and a path tracking control module. The transient tire cornering stiffness is estimated through an adaptive forgetting factor least squares method. Utilizing the estimated transient cornering stiffness values, the system identifies stable and unstable operating regions of the tire, facilitating an assessment of the current tire slip stability. By linearizing the tire model, the cornering stiffness state is predicted, generating desired tire forces that are subsequently fed into the path tracking controller. The controller dynamically adapts its configuration based on the tire’s operating region, thereby reducing the computational burden of the control solution. Upon receiving the desired tire forces, the controller modulates the unmanned forklift’s drive, braking, and steering systems to ensure the tires operate within a stable regime. Comprehensive simulations conducted using MATLAB/Simulink and PreScan, complemented by hardware-in-the-loop (HIL) testing, validate that the proposed path tracking control strategy reduces the peak lateral tracking error by approximately 25.4% compared to the NMPC benchmark on low-to-medium adhesion surfaces, thereby effectively enhancing lateral stability while ensuring real-time computational feasibility.