Research on control strategy of active suspension based on improved RBF neural network

Aiming at the problem that passive suspension cannot be adjusted independently to achieve good buffering effect on different road surfaces, a dynamic adaptive radial basis function (RBF) neural network PID controller is designed based on the two-degree-of-freedom hydro-pneumatic suspension model of 1/4 car body, which improves the traditional RBF neural network structure and relies on the self-learning behavior of neural network to dynamically adjust PID parameters, The simulation model was built in Matlab/Simulink, the simulation results showed that, under the condition of B-class and C-class uneven pavement input, compared with the multimode self-switching PID control, the root mean square error (RMSE) of tire dynamic load, body centroids acceleration and dynamic deflection of hydro-pneumatic suspension under the improved RBF-PID control are reduced by 42%, 64% and 3%, respectively, which verifies the superiority and effectiveness of the proposed control strategy.