Weeding Manipulator Exploiting Its Oscillatory Motion for Force Generation: Parameter Optimization of VDP Oscillator using Real-Coded Genetic Algorithm

摘要

This paper addresses parameter optimization of Van der Pol (VDP) oscillator driving a weeding manipulator, which exploits its own oscillatory motion to achieve efficient force generation. The author has confirmed the effectiveness of the exploitation in force generation using simulations. The simulation results have shown that a method using the VDP oscillator attains a superior performance to one using a sinusoidal wave function. In the simulations, however, the parameters of the VDP oscillator have been chosen after several attempts. Therefore, if the parameters are adjusted by an appropriate way, the performance of the force generation method will be improved. In this paper, real-coded genetic algorithm (RCGA), which is an optimization technique, is applied to the parameter optimization of the VDP oscillator because it is easily implementable for nonlinear systems. Simulation results show that RCGA succeeded in finding out sets of better parameters than ones used in the previous study.