A GA (Genetic Algorithm) based gait synthesis method for humanoid walking locomotion in both sagittal and frontal planes is proposed in this paper. The proposed method is able to optimize humanoid walking gait in terms of minimum energy consumption and maximum dynamic stable margin using ZMP (Zero Moment Point) criterion. Though there are dozens of parameters to be tuned in the GA based gait generation, only those greatly affecting the humanoid walking performance are considered. The method can generate an optimal combination of key parameters which results in the optimal performance of humanoid dynamic walking. In addition, an ANN (artificial neural network) is adopted to store and generalize the information of nearoptimal walking gaits obtained by GA optimization. Both simulation and real robot experiments are conducted to verify the effectiveness of our proposed method.

Keywords: Humanoid Robots, Walking Planning, Gait Synthesis, Genetic Algorithm, Neural Network, Trajectory Generation, ZMP.

( Full Paper, 1.17MB)

BibTex:

@ARTICLE{P1130548001,

AUTHOR = {Zhe Tang and Zengqi Sun and Changjiu Zhou},

TITLE = {GA Based Optimization for Humanoid Walking },

JOURNAL = {ICGST  International Journal on Automation, Robotics and Autonomous Systems},

YEAR = {2006},

VOLUME = {05},

ISSUE={II},

PAGES={1--10}

}

( Full Paper, 1.17MB)