Research On Visual-based Mobile Manipulator Motion Control

Visual-based mobile manipulator motion planning in general cluttered scene has been a very challenging problem in robotics research. An autonomous motion planning algorithm for floor-select and in-out of the elevator is designed in this paper,to guide the autonomous service robot taking the elevator autonomously in airport terminal.Firstly, the motion planning algorithm for floor-select is analyzed. The robot will operate the button before in-out of the elevator. The visual-based feedback motion control on the button is critically. The algorithm are constitute by fuzzy strategy path planning,outlined methodology in manipulator path planning and a real-time motion control algorithm based on Jacobian Matrix to approximate the button.Secondly, a new real-time path planning method which adjusts threshold adaptively to find feasible directions is developed for mobile robots based on laser range finder to solve the problem of restriction in path planning and motion control.The time in-out of the elevator must accurately. Otherwise, the robot will miss the time in-out of the elevator or have collision with it. The path planning method find feasible directions and velocity of the robot based on the information gathered by sensors. The failure that may arise during the process of taking elevator is analyzed together with the solution.Finally, the coordination of the robot control algorithm which guides the robot to take the elevator of the indoor environment is designed. Experiments are done on our humanoid service robot, and the results prove the validity of this control system.Visual-based motion planning provides a practicable way which allows robot move intelligently and autonomously. This paper provides pre-exploration for follow-up study on autonomous robot.