Parameters tuning of snake robots sidewinding gait using Bayesian optimization

In this paper, Bayesian optimization has been employed to maximize the forward velocity of a developed snake robot, when performing side-winding motion. Taking into account the snake body shape, the optimization has been performed in the high dimensional space of gait parameters by imposing appropriate bounds on the optimization parameters, which guarantees that each parameter to be evaluated will result a valid motion. Experimental results, which show 34 percent faster motion with necessary practical consideration are included to show the effectiveness of the proposed method.