Abstract:

Developing control schemes for different tasks that require extensive contact or interactions with the environment has been an important subject of research in robotics area. The main purpose of this thesis is to design a controller and achieve good position and force tracking performance when the robot makes contact with the environment.

The dynamic behaviors of the robot with flexible joint are discussed first, and then the solution of the control level is proposed based on singular perturbation theory. Impedance control is one of the main force-control schemes in literature; however, classical impedance control lacks of force tracking capability. It is more important to have the capability to track the specified desired force under uncertainties in the position of contact with the environment. This thesis aims to improve classical impedance control by proposing hybrid impedance control with an admittance compensator, which combines the advantages of the hybrid position/force control and the command-based compensator.

To verify the proposed controller, the 6-axis NTU arm model with flexible joints were constructed in ADAMS, and the entire control algorithm was built in MATLAB/Simulink. Simulation studies were shown to demonstrate tracking performance of the proposed control scheme under uncertainties in environment position. Experiments were conducted using a real-time control system with USB-to-CAN-Bus interface and the 6-axis NTU arm, which have been developed by our laboratory.