On developing a hybrid approach for kick optimization in humanoid robots

Abstract

The paper presents a hybrid approach for kick optimization in bipedal robots. It designs kicks using two different interpolation techniques, namely Hermite Spline and Bezier curves, and then optimizes their parameters using Particle Swarm Optimization while using inverse kinematics for execution. RoboCup Soccer Simulation 3D has been used as the test bed for the evaluation of the presented approach. The results show Hermite cubic spline to be a better interpolation scheme for this particular problem.