Robotics have been used in many arenas, such as manufacturing, medical, and space. Currently, the use of robotics is limited with respect to performance capabilities. Improving the performance of robotic mechanisms is a main research area. In this thesis, performance improvement is achieved through the approaches of robotic mechanism synthesis design, dynamic balance, and adaptive control.

A novel three degrees of freedom hybrid manipulator is designed. After discussing the advantages of this new type of hybrid manipulator, the kinematic and Jacobian matrix of this manipulator are analyzed. The kinematic performances, which include stiffness/compliance and workspace, are then analyzed and optimized, and the multiobjective optimization on the compliance and workspace is subsequently conducted. The dynamics of the proposed manipulator are analyzed based on the Lagrangian method.

When robotic mechanisms move, because the center of mass is not fixed and angular momentum is not constant, vibration is produced in the system. Dynamic balance is normally achieved by using counterweight(s), counter-rotation(s) or damping methods. However, the problem is that the whole system will become heavier and have more inertia. It is here proposed that dynamic balancing can be achieved through reconfiguration, rather than using counter-devices, so that the system will not gain any unwanted weight. After designing a balanced single leg, the legs will be combined to synthesize parallel mechanisms, i.e. first dynamically balance a single leg by the reconfiguration method (decomposition) and then combine the balanced legs to synthesize the whole parallel mechanism (integration).

As the mechanism is reconfigured, the control system has to be reconfigured accordingly. One way to address the control system reconfiguration is by breaking up the control functions into small functional modules, and from those modules assembling the control system. A hybrid controller for serial robotic manipulators is synthesized by combining a proportional–integral–derivative controller and a model reference adaptive controller in order to further improve the accuracy and joint convergence speed performance. The results show that the convergence speed for the hybrid controller is faster than that of the MRAC controller. The hybrid and MRAC controllers are both better than that of the PID controller. Experimental system is developed to model and verify the correctness of a reconfigured control system.