An Autonomous Navigation Approach Based on Birds–Eye View Semantic Maps

This paper presents autonomous navigation approach that an draws inspiration from birds’ navigation techniques, utilizing bird’s eye view semantic maps. These maps encapsulate high-level semantic information about the environment, such as roads, sidewalks, obstacles, and landmarks. By integrating this semantic understanding with a comprehensive aerial perspective, our approach aims to enhance navigation capabilities in dynamic environments. We leverage advanced computer vision techniques to construct detailed semantic maps, enabling autonomous agents to plan and execute maneuvers effectively. The proposed approach offers a promising solution for enhancing navigation robustness and adaptability in various real-world scenarios, including autonomous vehicles, robotics, and unmanned aerial systems.