Dynamics and Control of Multibody Cable-Driven Mechanisms with Application in Rehabilitation Robotics

With increasing demand for physical therapy in recent years, robotic systems have been proved to have great potential in improving the level of the delivered rehabilitation services both in quality and quantity and also providing huge savings in labor costs of. In this project a new cable-driven robotic cell for rehabilitation of human limbs is proposed and developed. This system has several advantages over the commercialized therapy robots, including reconfigurability and ability of handling redundancies. In this framework, the first challenge is to determine the number and configuration of the cables which guarantee the equilibrium of the system against an arbitrary force. The necessary and sufficient number of cables for single rigid body systems is well-known in the literature. However, since the human arm is a multibody system, a new theory was developed to determine the minimum total number of cables for a multibody as well as their possible distributions among the links of the multibody. In the second step, a method was proposed to obtain the boundaries of the workspace of the robot by Lagrangian formulation of dynamics of the multibody. Having the workspace, the final part of the thesis is on designing the control loop and its real-time implementation on a mechanical model of the human upper extremity. The control logic was designed in two levels: position control and compliance control. Position control is utilized for following a specified trajectory (representing an exercise for the patient's arm), while compliance control provides flexibility for deviating from that trajectory. The compliance control method used is impedance control in which the robot acts similar to a mass-spring-damper system. To achieve the exact stiffness required by impedance control, the inherent stiffness of the cable robot is formulated and taken into account by extending the theories proposed in the literature for stiffness of rigid-body cable robots for multibodies. Using impedance control enables us to perform different scenarios for training including teaching/playback and assistive/resistive exercising. The experimental results prove the effectiveness of the theories developed for dynamics and control of the multibody cable-driven robots.