基于改进模型预测控制的机器人自适应路径跟踪控制方法  

Adaptive model predictive control algorithm of robots based on fuzzy sliding mode

作　　者：应泽华 王立辉 顾炜琪 许宁徽 YING Zehua;WANG Lihui;GU Weiqi;XU Ninghui(Key Laboratory of Micro-inertial Instrument and Advanced Navigation Technology,Ministry of Education,School of Instrument Science and Engineering,Southeast University,Nanjing 210096,China)

机构地区：[1]东南大学仪器科学与工程学院微惯性仪表与先进导航技术教育部重点实验室,南京210096

出　　处：《中国惯性技术学报》2024年第11期1142-1150,共9页Journal of Chinese Inertial Technology

基　　金：国家重点研发计划(2022YFD2001503);江苏省重点研发计划(BE2022389)。

摘　　要：针对轮式机器人在地头转向路径跟踪误差累积的问题,提出一种基于自适应速度跟踪的改进模型预测控制算法。首先,设计基于模糊理论的自适应速度预测算法,研究不同横向偏差和预览道路曲率下的路径跟踪精度,得到纵向速度与横向偏差和道路曲率的模糊规则表。其次,对模型预测控制算法中的状态方程进行横纵向分离,设计滑模控制算法进行纵向速度跟踪控制。最后,在实际车辆平台上,针对U型和平滑型两种地头转向路径对所提算法进行实验验证。实验结果表明,相较于传统模型预测控制算法,所提算法路径跟踪精度提升28.9%。To address the issue of accumulated path tracking errors in wheeled robot headland turning,an adaptive speed-tracking improved model predictive control algorithm is proposed.Firstly,an adaptive speed prediction algorithm based on fuzzy logic is proposed to enhance tracking performance.Path tracking accuracy is evaluated across different lateral error and preview road curvatures,and a fuzzy rule table is formulated to describe the relationship between longitudinal speed,lateral error and road curvature.Secondly,based on the decoupling of the state equations in the model predictive control algorithm into lateral and longitudinal components,a sliding mode control algorithm is designed for speed tracking control.Finally,experiments are performed on a real vehicle platform using U-shaped and smooth-shaped headland turning paths.The results demonstrate a 28.9%improvement in path tracking accuracy and a 62.3%reduction in control cycle solving time compared to traditional model predictive control algorithms.

关 键 词：轮式机器人 速度预测 模糊控制理论 路径跟踪 模型预测控制算法 

分 类 号：U666.1[交通运输工程—船舶及航道工程]