Control of mobile manipulators

This dissertation presents a new class of control and coordination schemes for mobile manipulators. A mobile manipulator, which consists of a mobile platform and a robot arm, provides a new concept and direction in robot applications due to its capability of infinite motion in a large workspace and dexterous manipulation. Mobile manipulators have various applications in rescue and hazardous fields, battlefield, homes, offices, hospitals, etc. However, motion planning and control of such a system is challenging. This dissertation first presents the design of non-time based tracking controller and its application to nonholonomic mobile platforms. The non-time based tracking control is a key step to ensure the robustness of the mobile manipulator to unexpected obstacles and coordination errors. Then, singularity analysis and motion control of robot arms is discussed. Robot arm singularity avoidance is considered as one of the criteria for coordinating the motion of the mobile platform and the robot arm. Using a hybrid system approach, a singularity-free robot motion controller has also been developed for controlling the robot arm. Further, a unified dynamic model for the integrated mobile platform and on-board manipulator arm is developed. This provides an efficient and convenient framework to design a mobile manipulator controller as well as plan its motions. Combining the non-time based planning and control method with the nonlinear feedback technique, motion and force controllers at task level are designed. An online non-time coordinating scheme is developed to resolve the kinematic redundancy. Based on the decoupled and linearized model, this dissertation discusses an online non-time based coordination scheme. Force/position control along the same task direction is originally proposed in this dissertation. The proposed force/position control scheme is illustrated using a cart pushing control application, which requires both force control and motion control along the same task direction. An integrated task planning and control scheme while the mobile manipulator interacts with such moving objects is proposed. The proposed approaches are implemented and tested on a mobile manipulator consisting of a Nomadic XR4000 and a Puma 560 robot arm.