Bionic learning strategy applied on snake robot motion control

To solve the problems that the control parameters are difficult to be made certain and the non-ideal motion performance of snake robot, we develop a bionic autonomous strategy based on operant conditioning mechanism, which can realise autonomously and optimise adjust control parameters. The designed strategy is a 5-tuple and mainly contains: 1) state space; 2) action space; 3) state transfer function, meaning the consequence after action; 4) tendency function; 5) action selection strategy, adjusting the selected probability of each action. The designed bionic strategy is applied to motion control of snake robot. The experiment analysis is carried out combined with Webots robot simulation software. The results of the simulation experiment indicate that the skill of autonomous optimisation adjustment control parameters will gradually be enhanced with the learning proceeding.