Design and application of upper limb wearable exoskeleton robot arm system for brachial plexus patients

The brachial plexus is a large nerve group located in the armpit region, where the nerve roots coming out of the spinal cord form three large branches and connect. Damage to this nerve collection during birth or in periods results in the inability of the person to move his arm as he wishes. In this thesis, the design and control of a wearable exoskeleton robot arm for Brachial Plexus injuries was carried out. In the study, the wearable exoskeleton robot arm was designed in the Solidworks program. Necessary kinematic analyzes and angle limitations of the robot arm were performed in the Solidworks program. The designed exoskeleton robot arm was transferred to the MATLAB program with the "simmechanics" add-on and the necessary block diagram for simulation was created. PID controllers were added to the rotation points (DOF) (wrist, elbow shoulder horizontal, shoulder vertical) in the block diagram and the simulation of the system was provided with the entered P, I and D values. In the simulation, it has been observed that the wearable robot arm works stably. EMG muscle sensors were then placed on the user's forearm and upper arm. ADXL335 position and acceleration sensor has been added to the robot arm system to obtain position information in three axes. The exoskeleton robot arm system was operated, muscle data and position data were obtained. Using these data, the system was trained with the fuzzy logic algorithm in the MATLAB program and angle values were determined. A machine learning algorithm was added to the system with the same data and it was determined that the support vector machine algorithms gave the best result. The system controlled by both the fuzzy logic algorithm and support vector machines has been compared. In addition, each part of the system (hand, forearm, upper arm back and connections) was removed and mounted with a 3D printer. ESP32 microcontroller development board was used to control the system and it was run in real-time with EMG muscle sensors and position sensors.