Design, analysis and control of exoskeleton-type robotic arm for the disabled persons

In this study, an upper limb rehabilitation robot arm with 3 degrees of freedom was designed for disabled patients who have loss of limb function due to diseases such as stroke. The kinematic and dynamic analysis of the designed rehabilitation robot arm was examined and closed loop control method was applied for the position, speed and acceleration control of the robot arm. In the study, 3 different rehabilitation robot arm designs with 3 degrees of freedom were made, considering the design criteria such as human upper limb bone structure, movement structure and anthropometry, and the most suitable robot arm design was selected. The kinematic analysis of the designed robot arm was carried out with the Denavit-Hartenberg method to determine the robot arm end position with the given reference angle values. At the same time, using the Lagrangian method with the reference angles given, while the robot arm reaches the given angle values by following a certain trajectory, the dynamic analysis is performed by obtaining the torque values occurring in each joint. The results of these kinematic and dynamic analyzes of the designed rehabilitation robot arm were modeled in MATLAB/Simulink and MATLAB/Simscape environments and presented graphically, and these results obtained from two different applications were compared with each other and their accuracy was demonstrated. With the obtained torque values, suitable direct current motor parameters were selected for each joint. The transfer function of the direct current motor was created with the selected parameters. PID control is applied for position, speed and acceleration control of the designed robot arm during trajectory tracking. In the applied PID control method, the control parameters were obtained by trial-error method and Particle Swarm Optimization. Position, velocity and acceleration control in line with the parameters obtained with 2 different methods resulted in a more accurate result with the Particle Swarm Optimization method.