| 摘要: |
| 【目的】针对模糊纯追踪模型追踪误差大、响应速度慢的问题，分析收割机实际工作
环境和工作特征，将预测控制理论与模糊纯追踪模型融合，设计并实现一种应用于收割机
全自动作业的预适应路径追踪模型。【方法】该模型采用模糊控制理论中根据运动情况对
参数动态调节的方法，将纯追踪模型中的前视距离参数由原本的固定值替换为根据追踪误
差自适应的变化量，从而满足农业机械在农田复杂环境中的路径追踪需求。针对参数动态
调节不能立即生效问题，该模型采用预测控制理论，通过内置收割机运动学模型，在追踪
发生之前预测追踪结果，提前对追踪量进行校正，提高其追踪精度。【结果 / 结论】通过使
用 Matlab+Simulink 平台对基于预适应路径追踪模型的收割机 RAC 导航系统进行仿真和实际
测试，结果表明，改进后的导航系统相比基于模糊纯追踪模型的农机导航系统，追踪误差
下降了 91.9%。通过在搭载了 RAC 卫星接收机的收割机上部署该导航系统，经过测试，在
3 km/h 的速度下追踪误差为平均 0.63 m，满足低成本收割机全自动作业需求。 |
| 关键词: 收割机导航 路径追踪 模糊纯追踪模型 预测控制理论 |
| DOI：10.12105/j.issn.1672-0423.20200101 |
| 分类号: |
| 基金项目:国家重点研发计划课题（2017YFD0700403） |
|
| RAC navigation system of harvester based on pre-adaptive path tracking model |
|
Zhang Wenli1, Guo Yu1, Chen Kaizhen1, Wei Liguo2, Yuan Yanwei※2
|
|
1.Department of Informatics，Beijing University of Technology，Beijing 100124，China;2.Chinese Academy of Agricultural Mechanization Sciences，Beijing 100083，China
|
| Abstract: |
| ［Purpose］This paper presents a fuzzy pure tracking model to solve the problems
of large tracking error and slow response speed of tractors，by analyzing the actual working
environment and working characteristics of the harvester. The predictive control theory and the
fuzzy pure tracking model are merged to design and implement a pre adaptive path tracking model.
［Method］The model uses fuzzy control theory to dynamically adjust the parameters according
to the movement situation，and the forward-looking distance parameter in the pure tracking
model is replaced by the original fixed value with the adaptive variation according to the tracking
error，so as to meet the tracking requirements of agricultural machinery in the complex farmland
environment. Considering the problem that the dynamic adjustment of parameters cannot take effect
immediately，the model adopts predictive control theory. The built-in harvester kinematics model
predicts the tracking results before tracking occurs，corrects the tracking amount in advance，and
improves the tracking accuracy. The Matlab + Simulink platform was used to simulate and actually
test the RAC navigation system of the harvester based on the pre-adaptive path tracking model.
［Result/Conclusion］The results showed that the tracking error of the improved navigation
system was 91.9% lower than that of the agricultural navigation system based on the fuzzy pure
tracking model. By deploying the navigation system on the harvester equipped with a RAC satellite
receiver，the tracking error at the speed of 3 km/h is 0.63 m on average，which meets the
requirements of automatic operation of low-cost harvesters. |
| Key words: harvester navigation path tracking fuzzy pure pursuit model predictive control theory |
