基于离散鸽群优化滚动时域控制的无人机集群协同搜索

Cooperative search by unmanned aerial vehicle (UAV) swarms in complex three-dimensional environments is a task with widespread applications. However, traditional UAV area search methods typically only consider the movement of UAVs between adjacent grids in a grid map, neglecting the actual motion execution of the UAVs. This paper constructs a gridded target probability map and a pheromone map to describe dynamic environmental updates. It employs the Bayesian formula to update the probability of target existence. To address randomly moving dynamic targets in the environment, a pheromone revisit mechanism is introduced to guide UAVs toward grid areas that have not been visited for extended periods. The path optimization function is constructed by comprehensively considering the environmental cost, obstacle avoidance cost, and route cost of UAVs. To address the discrete optimization function and the tendency of pigeon-inspired optimization to fall into local optima, an improved discrete pigeon-inspired optimization (IDPIO) method is proposed. This method introduces pigeon-inspired optimization into the discrete domain and incorporates restart counting factors in the map and compass operators, periodically restarting the population to avoid local optima. Combined with the concept of receding horizon control, a UAV search path solving method based on improved discrete pigeon-inspired optimization with receding horizon control is designed. Simulation experiments and algorithmic comparisons demonstrate the effectiveness of this algorithm, which can maintain good search performance while ensuring safe flight in complex environments.