Negative Stiffness Control of Quasi-Zero Stiffness Air Suspension via Data-Driven Approach with Adaptive Fuzzy Neural Network Method

Zhixiang Ma,Xing Xu,Ju Xie,Xinwei Jiang,Feng Wang
DOI: https://doi.org/10.1007/s40815-022-01357-1
IF: 4.085
2022-07-24
International Journal of Fuzzy Systems
Abstract:To reduce effects of vibration of commercial vehicles during driving, this paper proposes a new semi-active quasi-zero stiffness air suspension to improve vibration isolation and a strategy for negative stiffness control, which guarantees commercial vehicles owning excellent vibration isolation performance under different driving conditions. The strategy is mainly composed of data driven approach and adaptive fuzzy neural network method. Firstly, support vector machine (SVM) is adopted to identify road condition. The load, speed and air pressure signals collected by sensors and indirectly obtained suspension dynamic deflection data are imported into the SVM model trained by data to obtain-specific road conditions. Then, with the obtained road condition, the optimal air pressure of the pneumatic linear actuators is searched, which decides the negative stiffness of the system and is later the target pressure. Adaptive fuzzy neural network which trained by large amount of data is used in the air pressure seeking process to make sure that the target air pressure of any driving condition is suitable. Finally, active disturbance rejection controller (ADRC) is applied to realize tracking control of target air pressure. Parameters of ADRC are set to adapt to variable driving conditions. The results of Hardware-in-Loop (HiL) tests indicate that the negative stiffness control strategy can effectively improve multi-objective performance of commercial vehicles under different driving conditions.
computer science, information systems,automation & control systems, artificial intelligence
What problem does this paper attempt to address?