Search papers, labs, and topics across Lattice.
The paper introduces RA-Nav, a risk-aware aerial robot navigation system that leverages semantic segmentation to classify obstacles into stationary, temporarily static, and dynamic categories. Risk estimation functions are designed for each category to construct a local risk map, enabling a risk-informed path search algorithm that balances path efficiency and safety. Simulations demonstrate RA-Nav's superior success rates compared to baselines in scenarios with sudden obstacle state transitions, validated with real-world data.
Aerial robots can now better navigate unpredictable environments thanks to a risk-aware system that anticipates sudden obstacle movements using semantic understanding.
Existing aerial robot navigation systems typically plan paths around static and dynamic obstacles, but fail to adapt when a static obstacle suddenly moves. Integrating environmental semantic awareness enables estimation of potential risks posed by suddenly moving obstacles. In this paper, we propose RA- Nav, a risk-aware navigation framework based on semantic segmentation. A lightweight multi-scale semantic segmentation network identifies obstacle categories in real time. These obstacles are further classified into three types: stationary, temporarily static, and dynamic. For each type, corresponding risk estimation functions are designed to enable real-time risk prediction, based on which a complete local risk map is constructed. Based on this map, the risk-informed path search algorithm is designed to guarantee planning that balances path efficiency and safety. Trajectory optimization is then applied to generate trajectories that are safe, smooth, and dynamically feasible. Comparative simulations demonstrate that RA-Nav achieves higher success rates than baselines in sudden obstacle state transition scenarios. Its effectiveness is further validated in simulations using real- world data.
