Objective:
This study aimed to develop a low-cost and practical person-position notification system to support social participation and safe mobility for visually impaired individuals and to verify its validity and reliability.
Methods:
Six healthy participants (three males and three females, aged 40s to 60s) were recruited. Participants were blindfolded and seated two meters away from a 32-inch monitor displaying videos of a person moving laterally or approaching the viewer. A computer vision system using a Raspberry Pi and camera module detected the person’s position and delivered tactile feedback via electrodes attached to the forearm through an electrical relay. Directional and distance recognition accuracy, total accuracy, reaction time, and subjective evaluation (5-point scale) were measured. A total of 45 randomized trials (9 position patterns: 3 directions × 3 distances, each repeated 5 times) were conducted.
Results:
The average accuracy for direction recognition was 91.7% (left: 91.7%, center: 93.3%, right: 90.0%), and for distance recognition, it was 85.0% (near: 88.3%, medium: 85.0%, far: 81.7%). Overall recognition accuracy was 85.6%. The average reaction time was 2.1 seconds, with the fastest responses for center direction and near distance. The highest subjective rating was for “clarity of stimulation” (4.2/5.0), while “comfort of stimulation” was rated lower (3.5/5.0).
Conclusion:
The developed system demonstrated high precision in transmitting direction and distance information and was confirmed to be effective as a support device for the visually impaired. While direction recognition showed high accuracy, distance recognition could benefit from further refinement. As a low-cost and high-performance technology, the system holds practical potential for clinical and daily use.
View full abstract