基于非线性模型预测控制的无人叉车路径跟踪  

Path tracking of unmanned forklifts based on nonlinear model predictive control

作　　者：张红琼 白国星 ZHANG Hongqiong;BAI Guoxing(Guizhou Electronic Technology College,Guian 550003,China;School of Mechanical Engineering,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]贵州电子科技职业学院,贵安550003 [2]北京科技大学机械工程学院,北京100083

出　　处：《现代制造工程》2025年第4期45-52,共8页Modern Manufacturing Engineering

基　　金：国家重点研发计划资助项目(2023YFC3806603);中国博士后科学基金资助项目(2022M710354)。

摘　　要：随着智能制造、智能仓储等行业的发展,无人叉车受到了越来越广泛的关注,但是目前针对无人叉车的研究工作主要集中于定位、感知和规划等技术,针对路径跟踪控制方面的研究尚不够成熟。目前无人叉车路径跟踪控制方面主要存在两方面的问题,一方面,现有控制方法均未考虑转向轮转角约束、转向轮转角速度约束等系统约束;另一方面,现有研究工作通常未采用货叉前端中点作为控制目标。为解决这些问题,首先,基于非完整约束原理建立了以货叉前端中点为控制点的无人叉车运动学模型;然后,在此基础上设计了基于非线性模型预测控制(Nonlinear Model Predictive Control, NMPC)的路径跟踪控制算法;最后,通过MATLAB软件进行了仿真测试并与Stanley控制等控制方法进行了比较。结果表明,提出的以货叉前端中点为控制点的NMPC路径跟踪控制器可以实现较高精度的路径跟踪,在不同工况的仿真结果中,横向误差幅值的最大值为0.054 9 m,收敛值不超过0.014 7 m,相比以前桥中点为控制点的NMPC路径跟踪控制器、Stanley控制器等,可以将横向误差幅值的最大值减小75.88%以上,收敛值减小88.34%以上;此外,所提出的控制器的实时性可以满足低速控制系统需求,每个控制周期内的计算时间不超过20 ms,小于控制周期的40%。With the development of intelligent manufacturing, intelligent warehousing, and other industries, unmanned forklifts have received more and more extensive attention, but the current research work on unmanned forklifts mainly focuses on positioning, sensing, planning, and other technologies, and the research on path tracking control is not mature enough. At present, there are two main problems in the path tracking control of unmanned forklifts. On the one hand, the existing control methods do not consider the system constraints such as steering wheel angle constraints and steering wheel angle speed constraints, and on the other hand, the existing research work usually does not use the midpoint of the front end of the forks as the control target. To solve these problems, first, a kinematic model of unmanned forklifts with the midpoint of the front end of the forks as the control point was established based on the principle of non-complete constraints, and then a path tracking control algorithm based on Nonlinear Model Predictive Control(NMPC) was designed, which was simulated and tested by MATLAB software and compared with Stanley control and other control methods. The proposed NMPC path tracking controller, which takes the midpoint of the front end of the forks as the control point, can realize high accuracy path tracking, and the maximum value of the lateral error amplitude was 0.054 9 m and the convergence value was not more than 0.014 7 m in the simulation results of different working conditions, which can reduce the maximum value of the lateral error amplitude by 75.88 % at least and convergence value by 88.34 % at least, compared with the NMPC path tracking controller with the midpoint of the front axle as the control point and the Stanley controller, etc. In addition, the real-time performance of the proposed controller can meet the requirements of low-speed control system, and the computation time in each control cycle was less than 20 ms, which was less than 40 % of the control cycle.

关 键 词：智能制造 无人叉车 路径跟踪 预测控制 控制点 

分 类 号：U471.1[机械工程—车辆工程]