Recognition of Arm Movements with Myo Armband

Abstract

This report is about the recognition of arm movements. It contains the introduction of the sensor and algorithm used to detect and classify the movements. The chosen device is called Myo Armband, which is wearable and could output the IMU and EMG signal of human arms. The report also discusses the method of interacting with the device and collecting data from it. There
are explanations of the principles of Support Vector Machine, K Nearest Neighbour and Neural Network. The implemented algorithm is Neural Network, and there are adjustments and discussions about different parameters on the Neural Network. In the later part of the report, the results will be listed and analysed. At the end of the report, improvements that remain to be
done will be listed as well as some new ideas of application on robotic arms.

Date of Award	2019
Original language	English
Supervisor	Uriel Martinez Hernandez (Supervisor)