Design and vision based control of a mobile manipulator

An autonomous mobile manipulator, a wheeled mobile platform carrying a manipulator arm, with a vision based target acquisition system can be able not only to find and maintain fixation on a moving target but also to manipulate or handle it while the system itself is in motion.In this context, a mobile manipulator is designed and constructed with its subsystems: electromechanical, electronics and computer parts. Resultant robotic system consists of differentially driven mobile platform and a planar manipulator with servo controllers and measurement circuits. A calibrated stereo rig is mounted on mobile platform. This vision system is to obtain task space feedback for mobile manipulator control. A computer is integrated to overall setup as a controller which exploits vision system data.Using vision based control methodology, two different control types for mobile platform and manipulator are applied. Mobile platform is controlled with a hybrid visual servo controller. To control instantaneous linear velocity of mobile platform, a position based visual servo algorithm is developed and angular velocity, it pivots, is controlled with image based visual servo control law. Manipulator motion is controlled via dynamic look and move control strategy, in which actuator references updated instantaneously according to acquired target position.Control of the mobile manipulator, first, is done by task sequencing. A task is separated in to two parts: Mobile platform moves till target is placed in manipulator workspace and in proper orientation that gripper (tracking) can hold, afterwards manipulator moves to hold target (handling).The second control strategy developed is focused on redundancy of mobile manipulator and its motion controlled via moving manipulator and mobile platform simultaneously