Researchers Build $88 Offline AI Smart Cane for Blind Users

It is a low-cost ($88) AI-integrated smart cane that runs fully offline, fusing an RGB camera with Time-of-Flight distance sensing to deliver vibrotactile and audio alerts about elevated obstacles and environmental context.

A conventional white cane cannot detect elevated hazards (like branches or signs) or provide semantic context, whereas the smart cane detects these risks and alerts the user via vibration and audio.

Yes. Its low cost and offline operation suit resource-constrained and low-connectivity environments, supporting accessibility programs across the region.

It achieved a macro-averaged F1 of 0.82 (precision 0.85, recall 0.81), with a mean end-to-end latency of 330 ms and 2.8 W peak power draw.