Westminster School and Harris Westminster Sixth Form – Years 12-13
In a sustainable world, everyone should feel secure. Today, we have many security devices – cameras, physical locks, passwords and more – but people have found ways to bypass most common, affordable ones. That’s why we’ve designed a new type of security – a dynamic ‘footprint’ that prevents possession theft.
The idea comes from gait recognition – the method of identifying someone by the way they walk. Security professionals usually do this from a distance through cameras. Our system would use an accelerometer to assess people’s unique walking styles. The system would then know whether the owner or an unknown party is holding a possession while walking.
Currently, our AI fits data to an accuracy of around 95% and can predict values correctly around 77% of the time. Ideally, these need to be higher, but as this is a prototype, with varying flaws in data collection, these values are expected.
We believe this security system is better than others because it’s easy to use. Your possessions, whether it’s a jacket or a phone, will know whether or not it’s you picking them up without you doing anything. So, if you leave your bag in a room and it’s stolen, it will alert you.