Movement detection by brainwave : local processing

Abstract

This project aims to develop a low-cost wearable system to detect the intention of movement for the living-alone elderlies by analysing the Electroencephalogram (hereinafter EEG) of the end-user, which could be used as a subsystem of a fall prevention system. A single channel EEG sensor with custom-made comb-shaped electrode records and collects the real-time raw brainwave signal from the end-user and communicates with the local computational unit through Bluetooth for signal processing and data analysis. A special brainwave signal that contains the movement intention of the end-user, namely Motion-related cortex potential (hereinafter MRCP), is to be detected. The detection result could be used to trigger other detection of the fall prevention system. The specific movement for this project is selected to be the sit-to-stand transition, as research in 2012 shows that inappropriate sit-to-stand transfers have been found related to 41% of all falls in the vulnerable elderly population [29]. In the current stage, the hardware was carefully chosen to mount the electrode on the top of the scalp, and it has been set up for repeated experiments to verify the relationship between EEG and sit-to-stand transitions. Experiments on different quantities of the electrodes were performed to explore the feasibility of using only a single electrode design. Besides, a machine learning algorithm has also been developed and integrated into this project to enhance the accuracy of detection.