System architecture diagram
Final prototype created for the fall detection device
Labelled accelerometer data for the right foot during a forward trip
Confusion matrix showing the final results from the system test
Fall-Proof
- Categories: Machine Learning, University, Electronics
- Tools used: MATLAB, Arduino, LSTM Neural Networks, 6-Axis IMU Sensors
- Project date: 2019-2020
- Project Report: Fall-Proof: A Wearable Fall Detection & Intervention Device for the Elderly
About
As my final-year project for university, I developed a wearable fall-detection device aimed at the elderly which uses machine learning with the inputs of slipper/belt-mounted sensors to detect and recognise different movement patterns. The aim was to integrate this in a wearable device along with a hip-mounted airbag system so it could recognise a fall in real-time and deploy the airbag to prevent injury to the user.