Global agriculture is confronting unprecedented pressures from population growth, diminishing arable land, and severe rural labor scarcity, necessitating the advancement of intelligent agricultural equipment. As a core component of precision farming, unmanned agricultural tractors demand highly accurate and robust path tracking control. However, conventional methods often fail to cope with unstructured terrain and dynamic wheel slip under real field conditions. This paper proposes an extended state observer (ESO)-based improved Stanley guidance law, which incorporates real-time sideslip angle observation, adaptive preview-based path curvature compensation, and a sliding mode heading controller. The ESO estimates lateral slip caused by varying soil conditions, while the modified Stanley law utilizes look-ahead path information to proactively adjust the desired heading angle during high-curvature turns. Both co-simulation in Matlab-Carsim and field experiments demonstrate that the proposed method significantly reduces lateral tracking error and overshoot, outperforming classical algorithms such as fuzzy Stanley and sliding mode controller, especially in U-turn scenarios and under low-adhesion conditions.

In order to improve the control accuracy and adaptability of drying control systems in complex paddy rice drying processes, the Gaussian process regression model predictive (GPR-MPC) drying process control strategy is designed. The strategy integrates the advantages of drying mathematical models and artificial intelligence algorithms. Firstly, based on the predicted moisture content of the drying mathematical model and the moisture content detection value, the Gaussian process regression is used to establish the model of moisture content prediction error. Secondly, the GPR-MPC control system is designed and simulation experiments are conducted to verify its effectiveness. Finally, the GPR-MPC intelligent control system of a grain dryer is designed and drying experiments are conducted with the grain dryer. The GPR-MPC intelligent control system testing experiment is conducted using a 15-ton cross-flow batch type recirculating grain dryer. The experimental result shows that the maximum, average, and variance of the grain moisture content control error are 0.4%, 0.165% and 0.114%, respectively. Compared to the MPC control system, the designed GPR-MPC intelligent control system has high prediction accuracy, small moisture content control error, and stable control system operation. The integration of drying mathematical models and artificial intelligence algorithms can effectively improve the drying effect and reduce dependence on data volume. This research is of great significance for promoting the development of intelligent drying technology.