Cooperative Search Strategy of Multi-UAVs Based on Hexagonal Grid

Aiming at the cooperative search problem in a wide-area unknown region, a cooperative search strategy of multi-UAVs based on hexagonal grid is proposed for the fairness of decision-making, This strategy is trained by reinforcement learning method, and adopts centralized training and distributed execution, and deals with the dynamic environment problem by fusing pheromone models. Simulation shows that the cooperative search algorithm has the functions of unknown region cooperative search and online threat avoidance replanning. Compared with the square grid training method, the reinforcement learning search algorithm based on the hexagonal grid after training expands the feasible directions in each decision planning cycle, which is more suitable for low-slow-small UAVs, and the performance is more stable after convergence.