A Navigation and Augmented Reality System for Visually Impaired People

Area of Science:

• Computer Science
• Human-Computer Interaction
• Artificial Intelligence

Background:

• Augmented reality (AR) and computer vision have advanced significantly, enabling smartphones to perform accurate 3D motion estimation.
• Existing systems for visually impaired navigation often rely on physical landmarks or deployed infrastructure, limiting their flexibility.
• There is a need for advanced assistive technologies that leverage current mobile capabilities for enhanced accessibility.

Purpose of the Study:

• To present ARIANNA+, an enhanced system for autonomous mobility of visually impaired individuals.
• To eliminate the reliance on physical infrastructure by utilizing virtual paths through ARKit.
• To improve environmental interaction by enabling object recognition and content access via machine learning.

Main Methods:

• Exploitation of ARKit for creating and following completely virtual paths.
• Integration of convolutional neural networks (CNNs) for object and building recognition.
• Utilization of common smartphones for processing, localization, and navigation.
• Provision of multi-modal feedback including haptic, speech, and sound.

Main Results:

• ARIANNA+ successfully enables navigation in both indoor and outdoor environments using virtual paths.
• The system provides context-aware information through object recognition, enhancing user interaction.
• Elimination of physical landmarks simplifies deployment and increases system applicability.
• Positive user experience is facilitated through intuitive haptic, speech, and sound guidance.

Conclusions:

• ARIANNA+ significantly enhances physical accessibility for visually impaired people by leveraging AR and AI.
• The system offers a flexible and robust solution for autonomous navigation without physical infrastructure.
• Future work could involve further personalization and integration with broader smart city initiatives.