VIDEO: Using AI to make being alone safer for those prone to falling

Posted on March 21, 2019

Michaela, Monet, and Carey have developed a proof-of-concept system that uses artificial intelligence techniques and smartphone accelerometer data to determine when a user has fallen. The system compares fall data found online with accelerometer data from an individual user's smart phone. A neural network, applied in Python, allows the system to learn over time to accurately distinguish the difference between a fall and simple, baseline movement in an individual user.

When the system decides the user has most likely fallen, it could automatically call for help, alerting emergency medical services or family. It's technology that could help those who are at risk of falling lead safer, more independent lives.

It's all part of the team's ELEC490 capstone design project.