Aiming at the problem of collaborative optimization of path planning and energy efficiency for hydrogen-powered unmanned aerial vehicle (UAV) in complex inspection tasks, this article constructs an intelligent decision-making framework based on Deep Reinforcement Learning (DRL). This framework takes DDPG algorithm as the core, integrates the constraints of high-efficiency interval operation of fuel cells and the dynamic replenishment judgment mechanism (DRJ), and builds an end-to-end path-energy consumption joint optimization model. Through the simulation test in three typical scenes, such as city, mountain and plain, the superiority of this method in multi-objective performance is verified. Compared with the traditional A* algorithm, the hydrogen consumption per unit mileage is reduced from 6.8g/km to 5.9g/km, the proportion of efficient operation time of fuel cells in mountain scenes is increased to 72.4%, the task interruption rate is reduced to zero when multi-machine scheduling, and the system availability rate is 91.3%. The results show that the proposed strategy can significantly improve energy efficiency while ensuring 98% task completion rate, which meets the engineering requirements of 100 consecutive trouble-free operations. This study provides a feasible technical path for hydrogen UAV to realize long-term, high-efficiency and autonomous inspection.
Han et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: