Joint torque control of a robotic manipulator requires a close dynamic description model involving the non negligible dynamics of the subsystems making up the system. The mathematical model for joint torque control of the robotic manipulator has been identified as one of the major sources of failures of commercial robots. The manipulator is basically made up of links connected by joints, and the torque that moves the links connected to a joint is produced by the joint actuator and also in practice, the control law is fed into the actuator inputs, therefore the actuator dynamics becomes non negligible dynamics in the dynamic modeling of the manipulator for robust joint torque control. Hence, a complete dynamic model of the manipulator which involves the link dynamics plus actuator dynamics was proposed. This paper focuses on the modeling of a 3DOF articulated manipulator based on independent joint (decentralized) scheme and the determination of the viscous damping coefficient for the joint torque control model. The independent joint model provides closer mathematical description of the manipulator and also enhances robust controller design. Joint damping coefficient B, was determined through experiment based on bode plot of the open loop gain. From the results, it was concluded that joints I and II achieved the best performance when B is 0.001N.m/rad /sec and 0.01N.m/rad /sec respectively.

The need to design a robot manipulator that can complete tasks satisfactorily in the presence of significant uncertainties brought about the continued advance research in robust system design. This paper focuses on the robustness analysis of a closed-loop controller for robot manipulator in real environment. The neglect of wide range of uncertainties and failure to study the fundamental behavioral responses during design stage of a control system result to the system failure in real environments. The robustness analysis studies these essential behavioral responses of a controlled system considering the significant uncertainties that exist in real environment in order to design a robust controlled system. It was concluded that the robot manipulator controlled system can only achieve robustness when it can maintain low sensitivities and zero steady state error, stable over the range of parameter variations and its performance continues to meet the specifications of the designer in the presence of wide set of uncertainties. Robustness and optimization of the robot manipulator can be achieved using closed-loop control technique. Bode plot can be used to ascertain the performance and robustness behavior of the controlled system in frequency domain. The disturbance rejection and disturbance rejection settling time describe how well and fast the controlled system can overcome disturbances.

The flexible joint robot is gaining more popularity in research and development because of its light weight and numerous applications. The major problems of the flexible joint robotic manipulator are poor tracking performance and instability. This work aims at improving the tracking performance and stability of the flexible joint robot based on the tracking error, damping time, overshoot and stability margins of the flexible joint model. To achieve this, a mixed synthesis method was applied. The mixed sensitivity synthesis is a robust control technique which uses adjustable weights to design a robust controller model which improves the performance and stability of a plant through loop shaping. From the results, the flexible joint model recorded damping time of infinity which is very high, gain margin of 22.8dB and very low phase margin of 3.21e-12deg. This means that the flexible joint model suffers from poor performance and it is unstable. The mixed synthesis controlled flexible joint model recorded low damping time of 0.993seconds, overshoot of 0%, tracking error of 0.0214dB, gain margin of 24.9dB and phase margin of 86.9degrees. This means that the mixed sensitivity synthesis controlled FJR achieved improved tracking performance and robust stability. The mixed synthesis control technique maintained negligible changes in damping time, tracking error and stability margins when the joint flexibility coefficient of the joint was varied to verify the robustness of the system. The work concludes that the flexible joint tracking performance and stability improvement was achieved using mixed sensitivity synthesis.