Synergistic control of hilly tractor slip: A dual-parameter approach utilizing BP neural network

Abstract

To address the coupling issue of longitudinal and lateral slip in tractors operating on lateral ramps in hilly and mountainous terrain, this study proposes a slip cooperative control method that enhances lateral stability. An optimal longitudinal slip index, constrained by lateral slip, is established as the primary control objective. A BP neural network-based model for predicting longitudinal and lateral slip is developed using training data collected from a self-built experimental platform under various conditions, including different lateral ramp angles, soil moisture levels, and traction forces. The study introduces a dual-parameter cooperative control architecture that combines “throttle opening-traction” control modes. This includes two operational modes: traction control and “throttle opening-traction” control. The controller parameters are optimized using an improved ITAE (Integral of Time Absolute Error) performance index, integrated with a particle swarm optimization algorithm (PSO-APID). Simulations conducted in Matlab/Simulink verify the effectiveness of this method under typical hilly terrain conditions. The results show that the proposed traction control mode, “throttle opening-traction” control mode, and their combined control mode effectively stabilize the longitudinal slip within the optimal range. Under the combined control mode, the mean lateral slip is reduced by 12.4% compared to the uncontrolled case. Additionally, under compound excitation, the PSO-APID controller demonstrates a shorter response time and smoother throttle and traction curves than conventional PSO-PID and classical PID algorithms, exhibiting superior dynamic characteristics. While maintaining the slip within the efficient range of 19%, the lateral slip is constrained to within 6%, providing theoretical support and a technical pathway for the intelligent control of agricultural equipment in complex terrains.