An Algorithm for Quadruped Robot's Inverse Kinematic Problems Based on PSO

Abstract

The inverse problem is an essential problem in robot kinematics. In order to deal with the inverse kinematic problems of hydraulic quadruped robot, the motion process of a single leg is modeled, simulated and analyzed. A new method based on particle swarm optimization (PSO) is proposed in this paper to overcome low computational efficiency, long calculation time and accumulative error. The proposed algorithm is attributed to: 1) limit the range of particles initialization; 2) select suitable compressibility factors; 3) employ linear interpolation method to improve the efficiency of the inverse algorithm. The simulation results demonstrate that the developed method can provide enhanced computational efficiency and significantly reduce the calculation time while ensuring the precision requirement, and can also eliminate the accumulative error.