ElderHelper - Elderly Fall Detection System

ElderHelper is an ongoing project to fulfill a life need and a gap in market. Elderly people falls more often than young and the trauma involved with the fall of elderly people is far worse to their health compared to the young's.

Majority of these falls happens inside their homes causing serious injuries.

Nowadays many elderly people live alone or stay alone for part of the daytime without near care.

Using a wearable device to detect falls and notify for assistance is a desirable solution, not only concerning the health of the elderly but for their family tranquility and public health systems economy.

There are several businesses on the market offering products or platforms as solutions for this problem, some of them uses wearable devices to monitor elderly but none of them can detects near 100% of accuracy the fall of an elderly.

The current project will use sensors, to monitor the movements of the elderly and an AI algorithm to distinguish between a fall and other movements.

We used AVNET SmartEdge Agile Brainium device to achieve this.

This device use an AI platform (Brainium) to easily train motion detection and notify a mobile device when this movement happens.

How to do

Purchase you AVNET SmartEdge Agile Brainium device from: https://www.newark.com/avnet/avtsea01a-m01a-s01a-c00/smartedge-agile-meta-sensor-iot/dp/84AC0532?st=smartedge

With device on hands, go to Brainium portal and...

• Create new Gateway (install gateway app on your mobile device or Raspberry PI)

• Add new device and associate to the Gateway

• Create new project

• Assign device to the project

• Create new AI Studio Motion Recognition Workspace

• Click on your new workspace to enter the workspace
• Create new Motion

• Fix device on a dummy

• Record sets of motions throwing dummy on floor n times selected on field "Set the number of motions" - throw dummy in different positions and movements

• Select all training sets and click Generate Model

• Apply best model to the device

• Go to project and configure device

• Add new AI Rule

• Choose IFTTT and the motion you created

• Go to https://ifttt.com/ and create account
• Add Brainiun IFTTT service named "Get all Brainium alerts on your device"

• Install and configure IFTTT app on your mobile device

• Simulate falls and you will receive an alert on your mobile device! Cool!