Mixed Integer Programming-Based Semiautonomous Step Climbing of a Snake Robot Considering Sensing Strategy

Abstract

We propose a control method for semiautonomous step climbing by a snake robot. Our method is based on mixed integer quadratic programming to generate the reference trajectory of the head of the snake robot online. One of the features of the method is that it determines suitable positions and time duration in which to sense the surroundings before approaching the step. Furthermore, constraints on velocity and acceleration are taken into account, so that the snake robot can securely follow the generated trajectory. Our method was applied to a snake robot equipped with a laser range finder, which is used for step detection. Experiments were performed to verify the efficacy of the method.