基于模糊预测控制的自动驾驶汽车运动路径仿真研究  

Research on motion path simulation of autonomous vehicle based on fuzzy predictive control

作　　者：张宏涛 刘文斌 信俊昌[3] 罗聪明 王天生 ZHANG Hongtao;LIU Wenbin;XIN Junchang;LUO Congming;WANG Tiansheng(Guangzhou Xinhua University,Guangzhou 510520,Guangdong,China;Guangzhou University,Guangzhou 510006,Guangdong,China;Northeastern University,Shenyang 110167,Liaoning,China;Nanfang College Guangzhou,Guangzhou 510970,Guangdong,China;Guangxi Vocational and Technical College of Communications,Nanning 530023,Guangxi,China)

机构地区：[1]广州新华学院,广东广州510520 [2]广州大学,广东广州510006 [3]东北大学,辽宁沈阳110167 [4]广州南方学院,广东广州510970 [5]广西交通职业技术学院,广西南宁530023

出　　处：《中国工程机械学报》2024年第6期728-732,738,共6页Chinese Journal of Construction Machinery

基　　金：国家自然科学基金资助项目(62072128)。

摘　　要：为了提高自动驾驶汽车控制系统响应速度,降低其运动路径跟踪误差,设计了模糊预测控制方法,并对汽车运动路径跟踪效果进行了仿真。给出了4轮汽车平面简图模型,建立了汽车动力学方程式。设计了汽车运动路径的离散线性化预测模型,引用模糊逻辑控制器对模型预测进行在线调整和优化,使汽车控制系统具有更好的抗外界干扰能力,从而得到汽车运动路径的最优轨迹。设定汽车目标运动路径,利用Matlab软件对汽车横向位移误差、横向加速度、航向角误差和航向角速度进行仿真,比较模型预测控制和模糊预测控制的输出效果。结果显示:采用模型预测控制方法,自动驾驶汽车响应速度较慢,横向位移和航向角跟踪误差较大,横向加速度和航向角速度变化较大。采用模糊预测控制方法,自动驾驶汽车响应速度较快,横向位移和航向角跟踪误差较小,横向加速度和航向角速度变化较小。自动驾驶汽车采用模糊预测控制方法,具有较好的响应速度和抵抗外界干扰的能力,从而提高自动驾驶汽车运动的稳定性。In order to improve the response speed of the autonomous vehicle control system and reduce the tracking error of its motion path,a fuzzy predictive control method was designed,and the effect of vehicle motion path tracking was simulated.A planar schematic model of a four-wheel vehicle was provided,and the vehicle dynamics equation was established.A discrete linearization prediction model for the vehicle motion path was designed,and a fuzzy logic controller was used to adjust and optimize the model prediction online,so that the vehicle control system has better resistance to external interference,and the optimal trajectory of the vehicle motion path can be obtained.Set the target motion path of the car,use Matlab software to simulate the lateral displacement error,lateral acceleration,heading angle error,and heading angle velocity of the car,and compare the output effects of model predictive control and fuzzy predictive control.The results show that with the model predictive control method,the response speed of the autonomous vehicle is slow,the tracking error of lateral displacement and heading angle is large,and the lateral acceleration and heading angular velocity change greatly.With the fuzzy predictive control method,the response speed of the autonomous vehicle is fast,the tracking error of lateral displacement and heading angle is small,and the changes of lateral acceleration and heading angular velocity are small.The autonomous vehicle adopts the fuzzy predictive control method,which has a better response speed and the ability to resist external interference,thus improving the stability of the autonomous vehicle movement.

关 键 词：自动驾驶汽车 模型预测控制 模糊控制 误差 仿真 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]