Fuzzy control of rider-motorcycle system using genetic algorithm and auto-tuning

This study investigates the stability control of a rider-motorcycle system based on fuzzy control in conjunction with both genetic algorithm (GA) and an auto-tuning method. The autotuning method, which entails tuning rules, is employed to on-line adjust output gains of fuzzy control. Although a GA has been used for fuzzy control in the literature, it has been confined to aiding membership functions' enaction. By contrast, this study employs a GA to determine optimal parameters in control rules for fuzzy control and in tuning rules for the auto-tuning method. Both computer simulation and experiment with regard to an inverted pendulum hinged to a rotating disk are carried out to represent the circular motion of the rider-motorcycle system, in which the inverted pendulum represents a rider's body in banking motion. The relation between riding speeds of the motorcycle and leaning angles of the rider is examined based on speed variations and Bode plots. Simulation and experimental results show the significant effect of the rider's banking angle on stability control.