无人驾驶电动拖拉机犁耕机组路径跟踪与耕深模糊控制策略研究

Fuzzy Control Strategy for Path Tracking-Tillage Depth of Unmanned Electric Tractor Ploughing Unit

摘要:

田间土壤比阻多变、地块凹凸不平造成拖拉机在犁耕工况下行驶速度和牵引阻力大幅波动，传统拖拉机无人作业系统难以满足路径跟踪精度和耕深均匀性的农艺要求。本文系统分析了无人驾驶电动拖拉机犁耕机组作业特性，提出了拖拉机犁耕机组路径跟踪与耕深模糊控制策略，包括模糊Stanley路径跟踪控制方法和预估补偿变论域模糊PID耕深控制方法，兼顾了拖拉机重负载犁耕工况下的犁耕深度与行驶速度，提高了电液悬挂系统的耕深均匀性和无人驾驶系统的直线性。为验证所提控制策略的有效性，搭建无人驾驶电动拖拉机犁耕机组试验平台，开展田间路径跟踪和犁耕试验。试验结果表明：本文提出的控制策略能适应复杂多变的田间环境，路径跟踪作业最大横向偏差不大于0.042m，犁耕作业横向偏差标准差不大于0.086m；行驶速度均值保持稳定，平均绝对误差为0.16km/h，平均相对误差为4.38%；耕深变异系数均值为9.29%，相较于模糊PID算法降低14.57%。本文提出的控制策略可根据不同土壤比阻、地块凹凸程度和牵引阻力，提高拖拉机犁耕作业的耕深均匀性和导航直线性，具有良好的适应性和鲁棒性，满足拖拉机犁耕作业农艺需求，为实现无人驾驶电动拖拉机多参数控制提供了参考。

Abstract:

Variable specific resistance of field soil and unevenness of the plot cause the tractor driving speed and traction resistance to fluctuate greatly under the ploughing condition, and the traditional unmanned tractor operation system is difficult to meet the agronomic requirements of path tracking accuracy and ploughing depth uniformity. The operation characteristics of unmanned electric tractor ploughing unit were systematically analyzed, and a fuzzy control strategy for tractor ploughing unit path tracking and ploughing depth was proposed, including the fuzzy Stanley path tracking control method and the predictive compensation variable-domain fuzzy PID ploughing depth control method, which took into account the ploughing depth and driving speed under the ploughing condition of the tractor with heavy loads, and improved the ploughing depth uniformity and the straightness of unmanned system of electro-hydraulic suspension. In order to verify the effectiveness of the proposed control strategy, a test platform for the ploughing unit of unmanned electric tractor was set up to carry out field path tracking and ploughing tests. The test results showed that the control strategy proposed can adapt to the complex and variable field environment, the maximum lateral deviation of the path tracking operation was no more than 0.042m, and the standard deviation of lateral deviation of ploughing operation was no more than 0.086m; the mean value of the driving speed remained stable, with an average absolute error of 0.16km/h and an average relative error of 4.38%;the mean value of the coefficient of variation of ploughing depth was 9.29%, which was 14.29% lower than that of the fuzzy PID algorithm. A control strategy proposed can enhance the ploughing depth uniformity and navigation linearity of tractor ploughing operation on the basis of different soil specific resistance, the degree of concavity and convexity of the plot, and the magnitude of traction resistance, with excellent adaptability and robustness to satisfy the agronomic requirements of tractor ploughing operation, which provided a reference for the implementation of the multiparameter control of the unmanned electric tractor.