Gait planning of concave transition for a wall-climbing robot

For a gait planning problem of a wall-climbing robot, an on-line adaptive algorithm to perform concave transition is presented in this paper. The wall-climbing robot has a biped-wheel hybrid locomotion mechanism, which can make itself transit between two different wall surfaces. Firstly, the mechanical structure and the gait of the robot are analyzed respectively. Then, a finite state machine for gait modeling is established. On this basis, the gait planning method is proposed by using the interpolation scheme and the BP neural net. The simulations and the experiments show that for the wall-climbing robot prototype system, the on-line gait planning algorithm is feasible, and can improve the adaptive control ability effectively.