Biogeography-based optimized adaptive neuro-fuzzy control of a nonlinear active suspension system

This paper presents an optimum network structure based on a BBO tuned adaptive neuro-fuzzy inference system (ANFIS) to control an active suspension system (ASS). The unsupervised learning via Biogeography-Based Optimization (BBO) algorithm is used to train the ANFIS network. The optimal proportional-integral-derivative controller tuned based on the LQR method is used to generate the training data set. ANFIS base on Fuzzy c-means (FCM) clustering algorithm is applied to approximate the relationships between the vehicle body (sprung mass) vertical input velocity and the actuator output force. BBO algorithm is used to optimize fuzzy c means clustering parameters. The numerical simulation results showed that the proposed optimized BBO-FCMANFIS based vehicle suspension system has better performance as compared with the optimal LQR-PID controller under uncertainties in both of reducing actuator energy consumption and the suppression of the vibration of the sprung mass acceleration, with a 43% and 9.5% reduction, respectively.