Abstract: As one of the most typical artificial environments, stairway presents a serious challenge to search and rescue robot in the ruins. To ensure that the robot can climb stairs with enough capability and stability, an online prediction control method for the stair-climbing of a robot, which is based on the analysis of the stair-climbing of a transformable tracked robot named as Amoeba-II, is proposed and developed. This method is not only applicable for the normal mobile robot, but also for the isomerism-modules robot with different mechanism modules. Based on the kinematics and quasi-static analysis of each stage which is determined by the motion planning for the stair-climbing procedure, the motion-coordination criterion and the crossing criterion for the robot has been established. Meanwhile, the interference-avoiding criterion is set up to avoid the interference between the non-tracked module of the robot and the stair. Based on the identification of the stair, the method can predict the stair-climbing capacity and the stability for combined motion-planning. The experiment for the stair-climbing of the robot is performed to certify the validity of the proposed online prediction method.

Key words: Motion planning, Online control method, Stair-climbing, Transformable tracked robot, Accurate reconstruction, Compressed sensing(CS), Over-complete dictionary, Sparse representation, Vibration signal