• Exoskeleton
• Hand
• High-Level Algorithm
• Intention Detection
• Wearable Robotics

The central topic of this thesis was the development of a high-level algorithm for the control of a wearable robotic device for the rehabilitation of the hand. The algorithm was created to be able to adapt to the characteristics of individual subjects. Eight healthy subjects offered their participation for a preliminary test. Each subject was asked to perform repetitions of 2 grasp movements under different starting and speed conditions. Electromyographic data extracted from finger muscles and position data extracted from the encoders mounted on the device were used to create an algorithm capable of detecting and classifying movements. For the movement detection, an automatic logic was proposed for the extraction of movement-specifics thresholds. For the movement classification, two independent classifiers for each type of data were created with the aim of distinguishing between the 2 grasp movements. Finally, the results obtained and a brief discussion on future clinical tests were listed.