System identification of the double inverted pendulum based on genetic algorithm

The double inverted pendulum system is a less-driven, multi-parameter, strongly coupled, highly nonlinear system. The precise mathematical model of the system is important to controller design. So establishing precise mathematical model is focused on in this paper. By the Lagrangian mechanics, the mathematical model of the double pendulum system is established initially. In order to identify the solution of the mathematical model with the actual measured data in experiment, the system identification is carried out, based on the improved genetic algorithm. It is taken as the identification destination to minimize the integration of absolute error between the theoretical data and the actual data. The improved genetic strategy is presented, which improves the ability of the global convergence and local search capabilities, enhancing the diversity of the population and avoiding the premature convergence effectively. Experiments show that the improved genetic algorithm can find the optimal value of system parameters effectively and precisely, which makes mathematical model describe the motion characteristics of the actual system more accurately.