农田土壤采样车点跟踪自动取土控制系统设计与试验

Design and Experiment of Point Tracking Automatic Sampling Control System for Farmland Soil Sampling Vehicle

摘要:

针对当前农田土壤采样作业劳动强度大、精准性差、自动化程度低等问题，基于电驱履带底盘设计了一种农田土壤采样车及其自主作业控制系统。对采样车整机结构组成、工作原理与控制系统方案进行了阐述，基于有限状态机方法设计了采样车作业流程控制策略。以采样车自主完成农田内一系列采样点作业任务为目标，设计了点跟踪控制流程，建立了采样车点跟踪模型，运用矢量场法设计了采样车点跟踪控制算法。上线性能试验结果表明，基于矢量场法设计的采样车点跟踪控制算法具有良好的纠偏性能，能够快速上线且以较小的横向偏差保持对目标路径的稳定跟踪，当初始横向偏差为2m、初始航向偏差为0°时，行驶速度0.3、0.6、0.9m/s上线时间分别为15.7、11.8、11.9s，上线距离分别为4.72、7.10、10.74m。连续点跟踪试验结果表明，行驶速度0.3、0.6、0.9m/s下采样车相对于各目标路径最大绝对横向偏差分别为0.081、0.107、0.210m，平均绝对横向偏差分别为0.018、0.022、0.050m，标准差分别为0.026、0.027、0.064m，采样车停止位置相对于各目标点的绝对距离偏差平均值分别为0.068、0.081、0.141m。土壤采样装置性能试验结果表明，各机构运行平稳，同一深度下采样耗时一致，样本质量变异系数较小。采样车能够根据所设计的控制流程与点跟踪控制算法实现对一系列采样目标点的准确连续跟踪，满足农田土壤采样自动化作业需求。

Abstract:

To tackle the problems in current farmland soil sampling operation, including high labor demand, poor sampling accuracy, low automation level etc., a farmland soil sampling vehicle with its automatic operation control system was designed based on an electric drive tracked chassis. The structure components, working principle and control system scheme were elaborated. The control strategy for the sampling vehicle operation process was designed based on the finite state machine method. To realize the goal of performing soil sampling at a series of target points in the field automatically, a continuous point tracking process was designed, and the point tracking model for the sampling vehicle was established. A point tracking control algorithm was developed based on the method of vector field. From the on-line performance test results, the point tracking algorithm designed based on vector field method had good deviation correction ability. When the initial lateral deviation was 2m and the initial heading deviation was 0°, the on-line time consumed by the sampling vehicle at 0.3m/s, 0.6m/s and 0.9m/s were 15.7s,11.8s and 11.9s, respectively, and the on-line distance travelled by the sampling vehicle at 0.3m/s, 0.6m/s and 0.9m/s were 4.72m, 7.10m and 10.74m, respectively. From the continuous point tracking test results, when the operation speed of the sampling vehicle was set at 0.3m/s, 0.6m/s and 0.9m/s, the maximum absolute lateral deviations with respect to the reference path were 0.081m, 0.107m and 0.210m, respectively, the mean absolute lateral deviations were 0.018m, 0.022m and 0.050m, respectively, the standard deviations were 0.026m, 0.027m and 0.064m, respectively, and the mean absolute distance errors of the sampling vehicle with respect to the target points were 0.068m, 0.081m and 0.141m, respectively. The performance test of the soil sampling device revealed that its mechanisms could work smoothly together, the consumed time at different locations for the same sampling depth was consistent, and the coefficient variation of sample masses was small. The sampling vehicle could realize accurate continuous point tracking to a series of target points under the control of designed control process and point tracking algorithm, which permitted that the requirements for automatic farmland soil sampling could be met.