Evolutionary optimized footstep planning for humanoid robot

This paper proposes a novel evolutionary optimized footstep planner for humanoid robot. Firstly, a footstep planner using univector field navigation method is proposed to provide a command state (CS) which is to be the input of modifiable walking pattern generator (MWPG) at each footstep. Then the MWPG generates the associated trajectories of every leg joint of humanoid robot at each footstep to follow the inputted CS. Secondly, the step length modification method which makes the humanoid robot step over low obstacles with minimum step length is proposed. Lastly, evolutionary optimization for the univector fields is presented. The univector fields are optimized by evolutionary algorithm (EA) considering objectives in navigation of humanoid robot for efficient navigation. The effectiveness of the proposed evolutionary optimized footstep planner is demonstrated through computer simulations for the simulation model of small-sized humanoid robot, HanSaRam-VIII (HSR-VIII).