TapLock uses tinyML on Arduino to protect your bike from thieves
Bike locks have not changed that much in the last few decades, even though our devices have gotten far smarter, so they seem in need of an update. Designed with this in mind, the TapLock is able to intelligently lock and unlock from either Bluetooth or taps on the enclosure. It uses a Nano 33 BLE Sense to detect tap patterns via an onboard accelerometer as well as BLE capabilities to communicate with the owner’s phone.
Because taps are not necessarily directional, the TapLock’s creators took an average of each accelerometer axis and charted the time between the peaks. After collecting a large sample of data, they used Edge Impulse to process the data and then train a model with an accuracy of 96.4%. This allows the owner to have some wiggle room when trying to lock or unlock the bike.
The team also developed a mobile app, which provides another way for the bike’s owner to lock or unlock the bike, along with some extra features too. After connecting to the TapLock, the app loads the previous state of the lock device and updates itself if needed. If the user wants to lock the bike, the app will send a “lock” command to the TapLock and store the current location to show on a map. This way the owner won’t forget where their bike is when trying to retrieve it.
Currently, the TapLock doesn’t have a physical locking mechanism, but the team states that one can be added and then electronically activated from one of the Nano 33 BLE Sense’s GPIO pins. You can see a demo of this project in the video below and read about it on Hackster.io.