Location
Engineering & Research Building
Start Date
3-12-2020 2:30 PM
End Date
3-12-2020 3:00 PM
Description
We designed and created an electric wheelchair that can be controlled via hand gestures using EMG data. Using a Myo armband, EMG data is collected and classified in real-time using a Neural Network. The raw EMG data is streamed to the Raspberry Pi using BLE where the Pi does signal processing and feature extraction. The features of the EMG signal are put through our Neural Network and the hand gesture is identified. The Neural Network works in real-time with an identification time of 375ms with a 96% accuracy. The network was trained to identify five unique hand gestures using databases created by us. We created a prototype wheelchair out of plywood and pvc that can be controlled via hand gestures. The system uses two DC motors, a Raspberry Pi, Arduino, and a motor controller in order to control and drive the wheelchair.
EMG Controlled Electric Wheelchair
Engineering & Research Building
We designed and created an electric wheelchair that can be controlled via hand gestures using EMG data. Using a Myo armband, EMG data is collected and classified in real-time using a Neural Network. The raw EMG data is streamed to the Raspberry Pi using BLE where the Pi does signal processing and feature extraction. The features of the EMG signal are put through our Neural Network and the hand gesture is identified. The Neural Network works in real-time with an identification time of 375ms with a 96% accuracy. The network was trained to identify five unique hand gestures using databases created by us. We created a prototype wheelchair out of plywood and pvc that can be controlled via hand gestures. The system uses two DC motors, a Raspberry Pi, Arduino, and a motor controller in order to control and drive the wheelchair.
