噪声与振动控制2025，Vol.45Issue(2)：163-170,203,9.DOI:10.3969/j.issn.1006-1355.2025.02.026
语义分割路面状态识别下汽车主动悬架EMPC控制
EMPC Control of Vehicle Active Suspension under Semantic Segmentation Road Condition Recognition
摘要
Abstract
In order to improve the smoothness of active suspension of the vehicles on random road and local pulse exci-tation road,an explicit model predict control(EMPC)method for active suspension was proposed based on semantic segmen-tation road surface recognition.A 2-DOF active suspension dynamics model was established;the DeepLabV3 semantic seg-mentation road recognition network was built based on empty space pyramid pooling,and the network was trained and veri-fied later.The EMPC control strategy of active suspension based on road recognition was designed.Then,the suspension dy-namic model was transformed into a prediction model,and the cost function and constraint conditions were determined.The optimal weighted weight of the cost function was matched according to the road recognition results.The system state param-eter regions were divided off-line,and the optimal control law of the system in each state region was solved.The designed control strategy was compared with passive suspension and linear-quadratic-Gaussian control(LQG)under random road and pulse road conditions.Compared with LQG control,the EMPC control strategy of active suspension based on road rec-ognition can improve the suspension performance under random road condition,and the adjustment time of suspension under pulse road condition is reduced by more than 20%,which effectively improves the ride comfort of the suspension.关键词
振动与波/路面识别/DeepLabV3/语义分割/主动悬架/EMPC控制Key words
vibration and wave/road surface recognition/DeepLabV3/semantic segmentation/active suspension/EMPC control分类
机械制造引用本文复制引用
寇发荣,王倩磊,张新乾,陈奕晓..语义分割路面状态识别下汽车主动悬架EMPC控制[J].噪声与振动控制,2025,45(2):163-170,203,9.基金项目
国家自然科学基金资助项目(51775426) （51775426）
西安市科技计划资助项目(21XJZZ0039) （21XJZZ0039）
咸阳市重点研发计划资助项目(2021ZDYF-GY-0027) （2021ZDYF-GY-0027）
