A robust adaptive controller for a nonholonomic mobile robot with unknown kinematic and dynamic parameters is proposed. A kinematic controller whose output is the input of the relevant dynamic controller is provided by using the concept of backstepping. An adaptive algorithm is developed in the kinematic controller to approximate the unknown kinematic parameters, and a simple single-layer neural network is used to express the highly nonlinear robot dynamics in terms of the known and unknown parameters. In order to attenuate the effects of the uncertainties and disturbances on tracking performance, a sliding mode control term is added to the dynamic controller. In the deterministic design of feedback controllers for the uncertain dynamic systems, upper bounds on the norm of the uncertainties are an important clue to guarantee the stability of the closed-loop system. However, sometimes these upper bounds may not be easily obtained because of the complexity of the structure of the uncertainties. Thereby, simple adaptation laws are proposed to approximate upper bounds on the norm of the uncertainties to address this problem. The stability of the proposed control system is shown through the Lyapunov method. Lastly, a design example for a mobile robot with two actuated wheels is provided and the feasibility of the controller is demonstrated by numerical simulations.
This paper describes an underwater 3500 m electric manipulator （named Huahai-4E, stands for four functions deep ocean electric manipulator in China）, which has been developed at underwater manipulation technology lab in Huazhong University of Science and Technology （HUST） for a test bed of studying of deep ocean manipulation technologies. The manipulator features modular integration joints, and layered architecture control system. The oil-filled, pressure-compensated joint is compactly designed and integrated of a permanent magnet （PM） brushless motor, a drive circuit, a harmonic gear and an angular feedback potentiometer. The underwater control system is based on a network and consisted of three embedded PC/104 computers which are used for servo control, task plan and target sensor respectively. They communicate through User Datagram Protocol （UDP） multicast communication in Vxworks OS. A supervisor PC with a virtual 3D GUI is fiber linked to underwater control system. Furthermore, the manipulator is equipped with a sensor system including a unique ultra-sonic probe array and an underwater camera. Autonomous grasp strategy based multi-sensor is studied. The results of watertight test in 40 MPa, joint＇s efficiency test and autonomous grasp experiments in tank are also presented.
