Robot gait synthesis using the scheme of human motions skills development

This work presents biologically inspired method of gait generation. Referring to the periodic signals generated by biological CPG, the coupled oscillators with correction functions are used to produce leg joint trajectories. The method of oscillators parameters evaluation using genetic algorithm is described. The correction functions are introduced. The gait formula suitable for real-time motion generation taking into account the positioning errors is given. Generated gait can be used as the basis for modifications to produce more complex motions. The problem of suitability of human gait for bipeds with body proportions different than that in human body is discussed.