Development of service hybrid mobile robot for climbing stairs and thresholds by switching wheel and leg gait

Abstract

We developed an autonomous omni-directional wheeled/4 legs(hybrid) mobile robot which has four driving omni-wheels arranged in a rectangular style and 4 legs of 2 DOF which like be a crab``s legs. This new locomotion mechanism that combines legs and robot body is proposed, and a prototype mobile robot that adopts the mechanism is described. We adopted Xscale processor as a main controller which is based on the ARM architecture enabling high-performance computing with low power consumption. Also by using Real-time Linux as OS, we constructed a Real-time control system. This mobile robot uses two modes for the mechanism, wheel mode and leg mode. Wheel mode is used to move in flat terrain, but leg mode is used to climb stair and threshold. In this thesis, only leg mode is focused and then main task of this robot is to arrive at the top floor of stair, to go over threshold autonomously. The motions of 4 legs are not just trivial because we also use the body of robot for more fast climbing and more stability. So the motion looks be abnormal, but robot is able to climb stair and go over threshold more fast. A trivial LRF can detect one dimensional distance. But we need the data of stair size which height and width of one step. Furthermore, we want to short the time of scanning as possible. Accordingly, we modified the only one direction of LRF by using mirror. A LRF can obtain another directional distance data by mirrors which has 3 dimensional angle. It necessarily has more complexity to obtain information of stair or threshold. But, it is worth sensing 3 dimesional information by just one scanning. The results of experiments are described to show the motion of 4 legs for climbing stair and going over threshold. Also joint thetas at each gait of legs, sensed data from LRF are showed.