【摘要】 环境探测与覆盖监控技术是环境保护、抗灾救援、资源勘探等领域的核心支撑，但其在复杂场景中面临信息分布不均、局部覆盖盲区等挑战。针对密集信息分布且需覆盖的未知单一区域，分析了多智能体协同覆盖控制中的通信优化问题。面对多智能体同步探测与覆盖中的环境感知低效、通信协同不足、动态适应性差问题，通过卷积神经网络处理环境图像、信号等数据，提高对复杂信息的感知能力，提供了精准的环境特征；利用图神经网络动态学习多智能体通信拓扑，提升通信效率与准确性；借助多层感知机的非线性映射能力，将信息转化为控制参数，使智能体能够灵活适应任务需求与环境变化。更多还原

【Abstract】 Environmental monitoring and coverage surveillance technology is a core enabler in fields such as environmental protection,disaster relief,and resource exploration. However,it faces challenges in complex scenarios,including uneven information distribution and local coverage blind spots. Focusing on unknown single areas with dense information distribution that require coverage,the problem of communication optimization in multi-agent cooperative coverage control is investigated in this paper. To tackle the issues of inefficient environmental perception,insufficient communication coordination,and poor dynamic adaptability in multi-agent synchronous exploration and coverage,convolutional neural networks(CNNs) are employed to process environmental data such as images and signals,enhancing the perception of complex information and providing accurate environmental features. Graph neural networks(GNNs) are utilized to dynamically learn the communication topology among multiple agents,improving communication efficiency and accuracy. Furthermore,the non-linear mapping capability of multi-layer perceptrons(MLPs) is leveraged to transform information into control parameters,enabling agents to flexibly adapt to task requirements and environmental changes.更多还原