Vision Assisted Navigation
Key Features of VAN:
- Seamless Indoor-Outdoor Visual Navigation: Supports uninterrupted navigation across both indoor and outdoor environments using only RGB images or videos—eliminating the need for GPS or specialized sensors indoors.
- Real-Time Object Detection and Path Planning: Employs deep learning models (e.g., YOLO, Faster R-CNN) to detect and track objects in real-time, enabling dynamic obstacle avoidance and intelligent path guidance.
- Device-Friendly and Scalable Deployment: Optimised for mobile and edge devices, the system supports applications like AR navigation, assistive tech, and wearable vision-based aids with minimal hardware requirements.
- Adaptive and Fusion-Ready Architecture: Capable of integrating with IMU, odometry, or sonar data, the system adapts to new environments through few-shot learning and continual updates, ensuring robustness and scalability