Navigation Scenario

Navigation scenario research focuses on enabling autonomous agents to efficiently and reliably reach destinations in diverse environments, ranging from indoor spaces to complex urban settings and even aerial flight. Current efforts concentrate on improving robustness to environmental uncertainty and sim-to-real transfer, employing techniques like Gaussian splatting for realistic simulation, large language models for high-level planning, and active inference for decision-making in ambiguous situations. These advancements are crucial for developing safer and more adaptable robots for various applications, including delivery, search and rescue, and autonomous driving.