Abstract: Considering that the traditional trial-and-error method of parameter setting for rotary inverted pendulum PD( Proportion Differentiation) controller has strong subjectivity and poor response ability, genetic algorithm is used to set parameters of PD controller so as to conduct model simulation and to ensure its operation on QUBE-Servo2 rotary inverted pendulum experiment system. The experiment shows that compared to the trial and error method, the PD controller parameters obtained by genetic algorithm further optimize the response performance of the system, and are not limited by subjective experience. The steady-state errors of the swing rod and swing arm are both within 0. 01 rad.

Key words: rotary inverted pendulum, mathematical model, proportion differentiation ( PD) control, genetic algorithm