A wearable mobility device for the blind using retina-inspired dynamic vision sensors

Area of Science:

• Assistive technology
• Sensory substitution
• Robotics and intelligent systems

Background:

• Visually impaired individuals face significant challenges in navigation and object avoidance.
• Existing assistive technologies often have limitations in real-time performance and user-friendliness.

Purpose of the Study:

• To propose and evaluate a prototype wearable mobility device for the blind.
• To assess the effectiveness of auditory-vision-substitution using dynamic vision sensors and event-based processing.
• To investigate different sound processing strategies for intuitive user interpretation.

Main Methods:

• Development of a wearable prototype employing two dynamic vision sensors.
• Utilizing event-based information processing for depth extraction.
• Conversion of 3D visual input to 3D auditory output via individualized head-related transfer functions.
• Evaluation of three distinct processing strategies with ten human subjects.

Main Results:

• The system demonstrated promising performance in assisting with navigation and object avoidance.
• Users achieved interpretable sound patterns with minimal training times (minutes).
• The device's design facilitates efficient real-time algorithms on a lightweight, hands-free platform with extended battery life.

Conclusions:

• The auditory-vision-substitution system shows significant potential as an effective mobility aid for the blind.
• The intuitive sound encoding and efficient processing enable rapid user adaptation.
• This wearable device offers a practical, low-power solution for enhancing independence and safety.