无人驾驶农机装备的模糊PI-LQR转向控制算法
Fuzzy PI-LQR Steering Control Algorithm for Unmanned Agricultural Equipment
姬江涛 1王启洲 1张玉成 2韩志豪1
作者信息
- 1. 河南科技大学 农业装备工程学院,河南 洛阳 471003
- 2. 中国科学院 智能农业机械装备工程实验室,北京 100086
- 折叠
摘要
无人驾驶农机装备转向系统处于小流量、小角度控制时会出现非线性和时滞,这会导致转向控制超调或者控制时滞,进而导致无法快速准确地跟踪到路径预瞄点.针对这个问题,本文提出了一种模糊比例积分-线性二次型调节器(PI-LQR)数据融合的内外环控制算法.该算法以模糊 PI控制为内环,基于转角误差和转角误差变化率,通过模糊逻辑运算自适应调节 PI参数,进行转向角度的闭环控制.以 LQR 控制为外环,基于农机装备 XY-横摆角控制,融合内环的转角误差,进一步提升无人驾驶农机装备转向的精准控制.Carsim 和Matlab/Simulink联合仿真表明:通过 100 m长的双移线路径下,最大横向误差为 0.18 m,最大横向摆角误差为0.067 rad.实车测试表明:通过 100 m长的双移线路径下,最大横向误差 0.26 m,相较于 PID 纯跟踪控制算法,最大横向误差减少了 0.19 m.
Abstract
The steering system of unmanned agricultural machinery exhibits the problems of nonlinearity and time delay when operating at low flow and small angle control.This phenomenon can lead to steering control overshoot or delay,resulting in the inability to accurately track the path preview point quickly.To address this issue,an inner-outer loop control algorithm based on fuzzy proportional-integral-linear quadratic regulator(PI-LQR)data fusion was proposed.The algorithm employs fuzzy PI control as the inner loop,adapting PI parameters through fuzzy logic operations based on steering angle error and its rate of change for closed-loop control of steering angle.The algorithm utilizes LQR control as the outer loop,incorporating the steering angle error from the inner loop based on agricultural machinery XY lateral angle control to further enhance the precise control of steering for unmanned agricultural machinery.Joint simulations using Carsim and Matlab/Simulink demonstrate that for a 100 m double lane change path,the maximum lateral error is less than 0.18 m,and the maximum yaw error is less than 0.067 rad.Real vehicle tests indicate that the maximum lateral error is less than 0.26 m for a 100 m double lane change path.Compared with the PID pure tracking control algorithm,the maximum lateral error is reduced by 0.19 m.
关键词
无人驾驶/转向系统/模糊PI/LQR/内外环控制Key words
autonomous driving/steering system/fuzzy PI/LQR/inner and outer loop control引用本文复制引用
基金项目
国家重点研发计划项目(2023YFD2001100)
河南省重大科技专项项目(221100110800)
河南省重大科技专项项目(231100220200)
出版年
2024
国家科技期刊平台
NSTL
万方数据