Fixed-Time Sliding Mode-Based Adaptive Path Tracking Control of Maize Plant Protection Robot via Extreme Learning Machine

Abstract

During the maize middle and late periods, the soil between rows is soft and also involved with weeds and straw. When the plant protection robot (PPR) moves on the soil, there exists uncertain shear perturbation because of the shear action and pressure subsidence, leading to the difficulty of the controller design. In this work, we propose an adaptive path tracking control (PTC) considering disturbances for the PPR. The disturbance of PRR in contact with soil is first revealed according to Bekker pressure subsidence and Janosi shear models, through which the plant model of PPR system is established. Then, we propose an adaptive fixed-time sliding mode (AFTSM)-based PTC to achieve excellent path tracking performance, where an extreme learning machine (ELM) estimator is developed, releasing the requirement for bound derivations in the control design. Using the fixed-time control and the ELM techniques in the proposed control, a remarkable control performance is well ensured, i.e., high-accuracy tracking, fast convergence, and excellent robustness. Experimental studies on a PPR are executed for demonstrating the validity and good performance of the designed controller.