1308 IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, VOL. 16, NO. 6, NOVEMBER 2008 Adaptive Robust Output-Feedback Motion/Force Control of Electrically Driven Nonholonomic Mobile Manipulators Zhijun Li, Shuzhi Sam Ge, Fellow, IEEE, Martin Adams, Member, IEEE, and Wijerupage Sardha Wijesoma, Member, IEEE Abstract—In this brief, adaptive robust output-feedback force/ motion control strategies are presented for mobile manipulators under both holonomic and nonholonomic constraints in the pres- ence of uncertainties and disturbances. The controls are developed on structural knowledge of the dynamics of the robot and actuators and in conjunction with a linear observer. The proposed controls are robust not only to parametric uncertainty such as mass vari- ations but also to external ones such as disturbances. The system stability and the boundedness of tracking and observation errors are proven using Lyapunov stability synthesis. Simulation results validate that the states of the system converge to the desired trajec- tory, while the constraint force converges to the desired force. Index Terms—Actuators dynamics, motion/force control, non- holonomic mobile manipulators, output feedback. I. INTRODUCTION T HE MOBILE manipulator possesses a complex and strongly coupled dynamics of the mobile platform and the robotic arm. Tracking control of mobile manipulators in practical applications requires both the motion and constraint forces converge to their desired trajectories and constraint forces, respectively, in the presence of parametric uncertainty [3], [6]. With the assumption of known dynamics, much research has been carried out. In [1], nonlinear feedback control for the mo- bile manipulator was developed to compensate for the dynamic interaction between the mobile platform and the arm to achieve tracking performance. In [2], coordination and control of mobile manipulators were presented with two basic task-oriented con- trols: end-effector task control and platform self posture control. In [6], force/position control of the end-effector for mobile ma- nipulators was developed using nonlinear feedback linearization and decoupling dynamics. To solve for the unknown parameters, adaptive schemes were investigated to deal with mobile manipulators with unknown in- ertia parameters and disturbances. In [3], adaptive trajectory/ Manuscript received July 24, 2007. Manuscript received in final form De- cember 10, 2007. First published June 10, 2008; current version published Oc- tober 22, 2008. Recommended by Associate Editor L. Villani. Z. Li was with the Department of Electrical and Computer Engineering, Na- tional University of Singapore, 117576 Singapore. He is now with the Depart- ment of Automation, Shanghai Jiao Tong University, 200240 Shanghai (e-mail: zjli@ieee.org). S. S. Ge is with the Social Robotics Lab, Interactive Digital Media Institute, and the Department of Electrical and Computer Engineering, National Univer- sity of Singapore, 117576 Singapore (e-mail: samge@nus.edu.sg). M. Adams and W. S. Wijesomais are with School of Electrical and Elec- tronics Engineering, Nanyang Technological University, 639798 Singapore (e-mail: eadams@ntu.edu.sg; eswwijesoma@ntu.edu.sg). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TCST.2008.917228 force tracking controls were investigated for mobile manipula- tors with unknown inertia parameters and disturbances. With the difficulty in dynamic modeling, adaptive neural network con- trols, a non-model-based approach [4], were developed for the motion control of mobile manipulators subject to kinematic con- straints in [5]. In these schemes, the controls are designed at kinematic level with velocity as input or dynamic level with torque as input, but the actuator dynamics are ignored. As demonstrated in [7], actuator dynamics constitute an important component of the complete robot dynamics, especially in the case of high-velocity movement and highly varying loads. Many control methods have therefore been developed to take into account the effects of actuator dynamics (see, for instance, [7]–[9]). However, most literature assumes that the actuator velocities are measurable [3], [5], which may deteriorate the control performance of these methods, since velocity measurements are often contaminated by a considerable amount of noise. Therefore, it is desired to achieve good control performance by using only joint position measurement. Moreover, in most research conducted on control of mobile manipulators, joint torques are the control inputs, while in reality the joints are driven by actuators (e.g., DC motors). Therefore, using actuator input voltages as control inputs and designing observer-controller structure for mobile manipulators with only the positions and the driving currents of actuators are more realistic. As such, actuator dynamics is combined with the mobile manipulator’s dynamics in this brief. This brief addresses adaptive robust output-feedback control of force/motion for a class of mobile manipulator systems elec- trically driven by DC motors with both holonomic and non- holonomic constraints in the parameter uncertainties and ex- ternal disturbances. Simulation results are described in detail that show the effectiveness of the proposed control. II. SYSTEM DESCRIPTION Consider an -degree-of-freedom (DOF) mobile manipulator mounted on a nonholonomic mobile base, the dynamics can be described as (1) where with denoting the general- ized coordinates for the mobile platform and denoting the coordinates of the manipulator, and . The sym- metric positive definite inertia matrix , the Cen- tripetal and Coriolis torques , the gravitational torque vector , the external disturbances , the known full rank input transformation matrix , 1063-6536/$25.00 © 2008 IEEE