For the last 2 years, Bluecover has been developing a new tracking and fitness device for golfing. By monitoring the golfer’s wrist, this device is able to automatically detect swings and can be used to track golf shots.
Our challenge was to miniaturise the electronics, to ensure minimal power consumption, and to integrate machine learning techniques in a bracelet that should be attractive and practical for golfing. The result is a wearable device that connects to a mobile application and supports intelligent classification.

Wearable device

The wristband’s electronics and firmware have evolved over a series of prototypes that were continuously improved over the development period. The evolution of the technology allowed us to embed increased functionality while reducing the physical dimensions of the device, with no compromise in efficiency. The wristband has become smaller, and wearing it has become more comfortable.
To accurately record full swing movements (which can reach accelerations of up to 20g!) we make use of sensors that provide pioneering levels of accuracy. These can capture the movement of the shot at very high sampling frequencies (up to 125 samples per second) to ensure the best possible resolution during the downswing (~250ms).
Energy efficiency was another concern. Our commitment to Bluetooth Low Energy (BLE) communications and the intensive use of a sleep mode optimise the wristband’s power consumption.

The main function of the wristband is to monitor the golfer’s wrist movements and to send the data to the mobile application for processing, where the user can view and analyse their swings in real time.

Mobile application

Swing Tempo is the mobile application for the driving range that receives and processes data from the Trueshot Wristband. The application currently provides the logic to detect golf swings (full swing, pitching and putting), is able to extract swing metrics such as timings, acceleration and angular rotation on shot, as well as estimate the putt distance and the club face speed for full-swings.

If the swing is correctly executed, the application will recognise it and extract metrics that can be analysed in real time (e.g. a swing tempo lower than 3 indicates that your backswing is too fast, so you should slow it down). On the other hand, if the swing is not correctly executed, the application will not recognise it, and you will know that you are doing something wrong.
The swing metrics stored in Swing Tempo can be used to keep track of your evolution or you can keep motivated in improving your game by seeing how you stack up with your friends.

The application will soon feature the ability to compare metrics with Scratch Players and provide tips for training improvement.

Machine Learning

The detection of golf swings is a smart feature that uses an automatic learning process based on recordings taken from golfers of different handicaps, genders and ages. The shots in this large set of recordings, which features full-swing, pitching, chipping and putting, were collected under supervision at driving range (depicted below) and then classified and saved to a database. The database, which currently features over 1000 swings, is a key component in the automatic learning and is being continuously expanded in order to achieve the best possible results.

The Trueshot wristband and Swing Tempo application are interoperable but commercially independent. The Wristband v3 (WB3) will soon be available for pre-order. The wristband was optimised for golfing but can be used to record and analyse movements in other sports or sectors. Swing Tempo is exclusively dedicated to golfing. It is available on Google Play, and is being continuously updated to support new metrics and features on top of WB3.