关键词: 无人驾驶, 电动拖拉机, 路径跟踪, 自适应模型, 权重策略

Abstract: Both pure tracking algorithm and Stanley algorithm are geometric tracking methods based on the control of vehicle front wheel angle, which are simple, direct, less control parameters and easy to implement, and are widely used. However, in the operation of unmanned tractor, the path tracking often fails to achieve the ideal effect due to the many influencing factors of farmland operation. In this paper, according to the characteristics of pure tracking algorithm and Stanley algorithm, an adaptive tracking algorithm is proposed, and the control model of the adaptive tracking algorithm is built. Through simulation analysis and optimization of the weight parameters, the optimal weight strategy is obtained. Comparative simulation tests were carried out under different soil environments, different velocities and different positions of the center of gravity. The simulation results show that the adaptive algorithm can realize the adaptive control under different working conditions. The positioning signal was sent to the vehicle integrated navigation receiver by the Chihiro base station, and the precise positioning information of the whole machine was obtained. The coordinates of the test plot were input into the navigation system, and the industrial computer was used to plan a reasonable operation path for the field experiment. The field experiments show that the adaptive algorithm can achieve the mean lateral deviation of 0.03 m in sowing operation and 0.11 m in field head turning. The mean lateral deviation of 22 rows of operations is within 0.05 m, which meets the requirements of operation accuracy.

Key words: driverless, electric tractor, path tracking, adaptive model, weight strategy