Lightweight Path-Aware Frontier Reordering for LiDAR UAV Exploration

Duc Manh Do and other authors published a paper proposing a lightweight path-aware frontier reordering method for LiDAR UAV autonomous exploration. The method first uses a lightweight utility function to retain a small number of top-N frontier candidates, then uses time-constrained RRT* planning to estimate the feasible path cost of each candidate and reorders them accordingly; the path of the selected frontier is also reused for navigation to reduce repeated global planning. The authors claim that in experiments in three warehouse-like environments, the method consistently produces shorter exploration trajectories than traditional frontier baselines, reducing travel distance by up to 14.1%; in two of the environments, exploration time is shortened by up to 9.5%, while in the third environment it is comparable to the traditional method. The authors also claim that its map detail is comparable to the traditional OctoMap-based baseline.

AI 根据报道生成 · 2 days ago更新