AR Indoor Navigation with Real Time Obstacle Detection

Abstract

Navigating indoor environments can be challenging. Traditional navigation aids often lack the precision and reliability required for safe and efficient indoor navigation. This project addresses this issue by developing an augmented reality (AR) indoor navigation system with real-time obstacle detection. The primary aim is to enhance the precision and dependability of indoor navigation systems, improving safety and user experience in indoor environments. The solution incorporates advanced obstacle detection, optimized sensor fusion methods, and robust algorithms. Leveraging machine learning and computer vision, the system detects obstacles in real-time and provides users with dynamic navigation guidance. The implementation utilizes technologies such as YoloV8 for Computer vision integration, Unity Engine for AR development, and Unity Authentication Services (also Firebase) for backend services. Test results demonstrate the effectiveness of the system in providing accurate navigation guidance and obstacle detection. Metrics such as obstacle detection accuracy, navigation responsiveness, and user satisfaction were evaluated to assess system performance. The AR indoor navigation system shows promise in enhancing spatial awareness and improving navigation experiences for individuals with visual impairments or navigating unfamiliar indoor environments.