A Simulation and Training Platform for Remote-Sighted Assistance
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.
View abstract on PubMed
Summary
This summary is machine-generated.Remote-sighted assistance (RSA) training using a virtual platform significantly improves agent performance. Trained agents guiding visually impaired people (VIPs) reduced guidance time by 32% and environmental contacts by 57%.
Area Of Science
- Human-Computer Interaction
- Assistive Technology
- Virtual Reality Training
Background
- Remote-sighted assistance (RSA) technology aids visually impaired people (VIPs) through remote agents.
- Video latency and instruction execution deviations challenge RSA effectiveness.
- Optimizing agent guidance under these conditions is crucial for VIP support.
Purpose Of The Study
- To develop and evaluate a virtual training platform for RSA agents.
- To enhance remote-sighted agent ability in guiding VIPs effectively.
- To improve mastery of tactile vibration commands for navigation assistance.
Main Methods
- Utilized Unreal Engine to create a realistic virtual training environment for RSA.
- Simulated VIPs executing instructions within the virtual environment.
- Trained remote-sighted agents on a set of encoded tactile vibration commands.
Main Results
- Agents trained on the platform demonstrated a 32.09% reduction in average guidance time.
- Trained agents showed a 57.57% decrease in average environmental contacts compared to novices.
- The platform facilitated effective training for guiding VIPs through complex paths.
Conclusions
- The virtual training platform offers a simple and effective solution for RSA agent development.
- Enhanced agent performance leads to reduced travel time and fewer environmental interactions for VIPs.
- This approach improves the overall effectiveness of remote-sighted assistance for the visually impaired.
These videos have been matched automatically. Contact us if they are not relevant.
These videos have been matched automatically. Contact us if they are not relevant.