Abstract: In the anti-rollover control of bus, there are various unknown nonlinear disturbances and parameter perturbations in actual vehicle system modeling process, so it is difficult to establish an accurate vehicle model, and there is problem of big chattering in standard sliding mode control (SMC). The RBF-ADSMC (radial basis function-adaptive sliding mode control, RBF-ADSMC) algorithm is proposed in this paper. Firstly, the radial basis function (RBF) neural network controller is used to estimate various unknown disturbance items and parameter perturbation items in vehicle modeling process. Then, the RBF neural network is used to adaptively adjust the key parameters of the standard SMC. Finally, the electronically controlled pneumatic hardware in the loop test bench is built, and the control algorithm is verified on the hardware in the loop test bench. The test results show that the RBF-ADSMC algorithm has good control effect and can meet the bus rollover control requirements. Compared with the SMC algorithm, the RBF-ADSMC algorithm can reduce the roll angle and lateral acceleration of the bus and improve the anti-rollover control effect of the bus

Key words: vehicle engineering, anti-rollover control, adaptive, RBF neural network, sliding mode control