// the challenge
The client wanted to build smart glasses for blind people, able to detect obstacles through ultrasonic sensors and communicate over Bluetooth LE with a smartphone. The device required ultra-efficient firmware to preserve battery life, and a Machine Learning algorithm capable of classifying obstacles in real time directly on the edge device.
// what we did
- Firmware on a Nordic nRF SoC with power consumption optimisation
- Bluetooth LE stack for communication with the mobile app
- Development of an on-device ML algorithm for obstacle recognition and classification
- Integration of ultrasonic sensors and real-time interrupt handling
- Support for the industrialisation of the wearable product
// application field
Engineering · FW and algorithm development
// gallery
