Robust Planning System for Fast Autonomous Flight in Complex Unknown Environment Using Sparse Directed Frontier Points

Path planning is one of the key parts unmanned aerial vehicle (UAV) fast autonomous flight in cluttered environments. However, it remains a challenge to efficiently generate high-quality trajectory for tasks with high success rate. In this paper, robust framework proposed, which can stably support complex unknown environments limited onboard computing resources. Firstly, we propose directed frontier point information structure (DFP), roughly capture explored environment. The direction local planner be evaluated and rectified based on DFP avoid falling into traps cost. Secondly, an adaptive fusion replanning method designed by incorporating two optimization methods different characteristics, both take advantage while avoiding disadvantages as much possible, but also adjust focus according actual situation improve rate method. Finally, sufficient comparison evaluation experiments simulation are presented. Experimental results show proposed has better performance, especially terms adaptability robustness, compared typical state-of-the-art scenarios. Moreover, system integrated fully quadrotor, effectiveness further using quadrotor real-world