融合LSTM与模糊控制的智能车辆横摆稳定性控制研究  

Research on yaw stability control of intelligent vehicles integrating LSTM prediction and fuzzy control

作　　者：窦强宾 王洪亮[1] 皮大伟[1] 王显会[1] 王尔烈[1] DOU Qiangbin;WANG Hongliang;PI Dawei;WANG Xianhui;WANG Erlie(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,Jiangsu,China)

机构地区：[1]南京理工大学机械工程学院,江苏省南京市210094

出　　处：《农业装备与车辆工程》2023年第12期69-74,共6页Agricultural Equipment & Vehicle Engineering

摘　　要：运动控制作为智能驾驶车辆的核心研究领域之一,单独的纵向控制和横向控制已经无法满足多变的道路条件要求,所以需要在考虑车辆行驶稳定性的前提下设计综合的控制策略。为了提高智能车辆在复杂条件下的稳定性,提出了一种将LSTM网络与模糊控制相结合的智能车辆横摆稳定性控制策略。通过训练LSTM完成车辆参数之间的非线性映射关系,使状态预测输出尽可能接近实际值。为了保证控制系统的实时性,利用LSTM网络通过已知车辆状态对车辆下一时刻的状态进行预测,通过预测状态对车辆的稳定性进行估计。当车辆处于不稳定状态时,以质心滑移角和横摆角速度为优化约束,采用基于模糊规则的自适应控制策略将车辆的状态误差估计转化为控制变量。仿真实验验证了所提出的控制策略能够保证复杂工况下的路径跟踪精度,提高车辆的稳定性。As one of the key research fields of intelligent driving vehicle,motion control can not satisfy the requirements of changeable road conditions by longitudinal control and lateral control alone,therefore,it is necessary to design a comprehensive control strategy under the premise of considering the vehicle driving stability.This paper proposes an intelligent vehicle yaw stability control strategy that integrates the LSTM network and fuzzy control.The nonlinear mapping relationship between vehicle parameters was completed by training the LSTM so that the predicted state output was as close to the actual value as possible.To ensure the real-time performance of the control system,the LSTM network was used to predict the state of the vehicle at the next moment through the known state of the vehicle,and the predicted state was used to estimate the stability of the vehicle.When the vehicle was unstable,an optimization constraint was constructed with the center of mass slip angle and yaw rate,and an adaptive strategy based on fuzzy rules was used to convert the estimated state error of the vehicle into a control variable.It was verified by simulation experiments that the proposed control strategy could ensure the accuracy of path tracking and improve the stability of the vehicle under complex working conditions.

关 键 词：智能车辆 轨迹跟踪 LSTM 模糊控制 横摆稳定性控制 

分 类 号：U463.6[机械工程—车辆工程]