Fall Alert

My name is Freddie Howells; I’m 14 years old and I’m from South Wales. My project is called Fall Alert and is designed to help elderly people, particularly those with dementia. It is a wearable fall sensor that automatically sends an alert to the family or carers if the elderly person falls. The fall event can also trigger an optional autonomous robot with a webcam interface, or a wall-mounted webcam system, to search the house for the location of the elderly person.

I was inspired to create Fall Alert by my Great Aunt Pat, who’s 90 and lives alone at home with dementia. I came up with the idea for Fall Alert because her current system requires her to press a button if she falls, but because of her dementia, when she has fallen, she has forgotten to press her button and has been found several hours later, still on the floor, by her carers. To solve this problem, my system automatically detects if she has fallen, so that she doesn’t have to press a button.

My project uses a Raspberry Pi micro-computer as the main processor, which is connected to an 3-axis accelerometer. The accelerometer is used to monitor everyday movements. I looked at the features of a fall and created a unique fall algorithm that incorporates all the aspects of a fall, so an alert is triggered if a fall event is recognised. This can then launch the optional robot or camera system, which uses computer vision and deep neural networks to search for a fallen person.

By combining FallAlert with GPS functions, there is also the option to add a tracking and alert feature if the wearer wanders away from home.

I am now working on reducing the size of the device into a compact bracelet-style unit and have designed a customisable bracelet using CAD software. I’m hoping to use crowdfunding to fund my own 3D printer so I can create my own prototypes for the bracelet.