Integrating Deep Planning-Based Object Detection with 3D-Depth Camera for Collision Avoidance in Indoor Robotics Navigation

This paper presents a framework for collision avoidance in indoor robotic navigation, integrating vision-based object detection and distance estimation, and rule-based decision-making. The system leverages YOLOv5 for real-time object detection and distance estimation, enabling accurate spatial awareness in dynamic environments. A key feature of the framework is a rule-based action mapping mechanism, which dynamically adjusts the robot's trajectory based on obstacle proximity and predefined safe distances. By seamlessly fusing visual and depth camera data, the proposed system enhances navigational safety and efficiency. The methodology is validated through simulations and real-world experiments, demonstrating its efficacy in collision avoidance and path optimization. This work contributes to the advancement of autonomous navigation systems, offering a scalable solution for complex multi-robot environments.