LucidLink^®

Everyday tasks such as recognizing a friend, finding a phone, reading a label, or understanding a room can require another person’s help. The challenge was to bring these needs together in a stable, hands-free wearable that feels direct and dependable.

The approach combined a Raspberry Pi 5, camera, microphone, speaker, and modular AI services. Voice commands provide the entry point, while each specialized service handles the relevant visual task and responds with clear speech.

• Chest-mounted format for a stable, hands-free camera view
• “Hey Lucid” wake word and natural spoken requests
• Dedicated services for faces, text, belongings, and scenes
• Spoken responses designed for fast, user-led interaction

LucidLink turns a simple voice prompt into useful environmental awareness. The system processes visual input locally or through specialist services, then uses speech synthesis to communicate the result back to the wearer.

LUCIDLINK CAPABILITIES

• Face recognition for familiar people and their relative position
• PaddleOCR-powered document reading and summaries
• Belonging recognition for saving and locating personal items
• Scene description for understanding people, places, and activity
• Help service for answering context-aware questions about a captured view
• Time service for hands-free date and time information
• Interrupt service that lets a new “Hey Lucid” request take priority at any time

LucidLink demonstrates a unified, multi-modal approach to assistive technology: a wearable device, voice-first interaction, and AI services working together to support greater autonomy in everyday environments.

“From the moment I slip on LucidLink and say ‘Hey Lucid,’ I feel like I have a trusted guide right beside me.”

Evaluation participant

Feedback documented in the capstone report