基于路面等级识别的车辆半主动悬架内外环控制
Semi-active Suspension Inner and Outer Ring Control of Vehicles Based on Pavement Grade Recognition
寇发荣 1郭杨娟 1刘朋涛 1门浩1
作者信息
- 1. 西安科技大学 机械工程学院,西安 710000
- 折叠
摘要
针对车辆在不同路面等级下对悬架动态性能与馈能特性需求不同的问题,提出一种基于RF-XGBoost路面等级识别算法的半主动悬架内外环控制策略.利用随机森林(Random Forest,RF)模型对极端梯度提升(Extreme Gradient Boosting,XGBoost)算法进行优化,搭建RF-XGBoost算法模型对路面等级进行识别.将路面等级与悬架控制策略相结合,设计外环为天地棚控制,内环为自适应滑模控制的内外环控制,实现非线性悬架的自适应控制.仿真结果表明,相比传统混合天地棚控制的悬架,内外环控制下的悬架在A级路面下簧载质量加速度降低15.52%,并实现50.4 W的振动能量回收,在B、C级路面下簧载质量加速度分别降低15.09%、16.72%,轮胎动载荷分别降低11.63%、11.42%,在D级路面下轮胎动载荷降低14.12%.台架试验的结果与仿真分析的结果基本一致,表明所设计的自适应内外环控制有效.
Abstract
Aiming at the problem that vehicles have different requirements of suspension dynamic performance and energy feed characteristics under different road surface grades,a semi-active suspension inner and outer ring control strategy based on the RF-XGBoost pavement level recognition algorithm is proposed.The Random Forest(RF)model is used to optimize the Extreme gradient boosting(XGBoost)algorithm,and the RF-XGBoost algorithm model is used to identify the pavement level.Combining the pavement grade with the suspension control strategy,the outer ring is designed to control the heaven and earth shed,and the inner ring is designed to control the adaptive sliding mode of the inner and outer ring,so as to realize the adaptive control of the suspension.Simulation results show that compared with the suspension controlled by the traditional hybrid shed,the spring load acceleration under the inner and outer ring control is reduced by 15.52%on the A-class pavement,and the vibration energy recovery of 50.4 W is achieved;on the pavements of B and C grades,the reed mass acceleration is reduced by 15.09%and 16.72%respectively,the dynamic load of the tire is reduced by 11.63%and 11.42%respectively;on the D-class pavement,the dynamic load of the tire is reduced by 14.12%.The results of the bench test are consistent with the results of the simulation analysis,indicating that the adaptive inner and outer ring control of the design is effective.
关键词
振动与波/路面识别/随机森林/XGBoost算法/混合天地棚控制/自适应滑模控制Key words
vibration and wave/pavement identification/random forest/XGBoost algorithm/mixed shed control/adaptive sliding mode control引用本文复制引用
基金项目
国家自然科学基金(51775426)
陕西省重点研发计划(2020GY-128)
出版年
2024
国家科技期刊平台
NSTL
万方数据