Affordable AI-integrated smart cane shows feasibility for offline, edge-native mobility assistance

Researchers describe a prototype smart cane integrating multimodal sensing and an INT8-quantized SSD MobileNet V1 model running on a low-cost Raspberry Pi Zero 2W, avoiding cloud connectivity to reduce cost and preserve privacy.

The system combines RGB vision with Time-of-Flight distance sensing and delivers distance-aware vibrotactile feedback and real-time audio alerts to assist visually impaired users navigating indoor environments.

In experiments across indoor mobility scenarios, the prototype achieved a macro-averaged F1-score of 0.82 (precision 0.85, recall 0.81), a mean end-to-end latency of 330 ms, and a peak power draw of 2.8 W.

A preliminary usability study with 12 participants reported a System Usability Scale score of 78.5 and indicated enhanced obstacle awareness, as measured by NASA-TLX workload assessments.

The authors argue the design is feasible for deployment in resource-constrained settings due to its offline operation, low hardware cost ($88), and edge-native architecture that isolates sensor acquisition, neural inference, and haptic feedback into independent processes with fail-safe sensing support.