改进模型预测控制的农机轨迹跟踪算法研究  

Research on Trajectory Tracking Algorithm of Agricultural Machinery Based on Model Predictive Control

作　　者：刘林 张华强[1] 秦昌礼 陈雨 苏庆华[3] LIU Lin;ZHANG Huaqiang;QIN Changli;CHEN Yu;SU Qinghua(School of Mechanical Engineering,Shandong University of Technology,Zibo 255049,China;Military Representative Office of Rocket Army in 211 Factory,Beijing 100076,China;School of Information Engineering,Beijing Wuzi University,Beijing 101149,China)

机构地区：[1]山东理工大学机械工程学院,山东淄博255049 [2]北京航天发射技术研究所,北京100076 [3]北京物资学院信息学院,北京101149

出　　处：《哈尔滨理工大学学报》2023年第1期29-36,共8页Journal of Harbin University of Science and Technology

基　　金：国家自然科学基金青年基金(61803035)。

摘　　要：针对无人驾驶农机进行轨迹跟踪时精度与稳定性较差的问题,提出了一种基于模型预测控制(MPC)的轨迹跟踪方法。首先,建立农机车辆的运动学模型,利用粒子滤波对农机状态量进行估计来提高农机的定位精度;其次,在设计MPC控制器时引入了梯度投影算法,该算法相比于传统有效集算法减少了迭代步数,具有更快的收敛速度,提高了农机进行跟踪控制时的计算效率;最后,在淄博市某无人农场进行了农机轨迹跟踪试验。结果表明:该方法可以实现良好的跟踪控制精度,符合精准农业的作业要求。In order to overcome the problem of low accuracy and stability when pilotless agricultural machinery performs path tracking,this paper proposed a trajectory tracking method based on model predictive control(MPC).Firstly,the kinematic model of agricultural machinery was established,and the state quantity of agricultural machinery was estimated by particle filter to improve the positioning accuracy of agricultural machinery.Secondly,the gradient projection method was introduced in the design of controller.Compared with the traditional active set method,the new method reduced the number of iterative steps and has a faster convergence speed,which can improve the computational efficiency of agricultural machinery tracking control.Finally,a path tracking experiment was carried out on the ecological farm in Zibo,Shandong Province.This algorithm can achieve preferable tracking control accuracy and meet the requirements for precision agriculture.

关 键 词：农机 轨迹跟踪 模型预测控制 投影梯度 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]