Spherical Gesture Control Interface for IoT Devices

This project was developed at BOCCI to create a universal IoT controller that seamlessly integrates sophisticated technology into beautifully crafted natural objects. The challenge was embedding capacitive sensors throughout a wooden or stone sphere while implementing neural network gesture recognition that could operate within strict low-power constraints.

The design brief was to develop a device which translates subtle hand gestures and touch patterns into commands through discovery-based learning; there would be no designators for orientation, or size of gesture, so it required the creation of a flexible, and complex recognition system.

To achieve this, I hired a small team of specialized contractors to address some of my knowledge gaps in neural network optimization and efficient data communication across multiple sensors while maintaining the low compute power requirements. The collaboration allowed me to focus on the integration challenges while ensuring the technical performance met BOCCI's exacting standards for both functionality and aesthetic refinement.