Abstract
This paper addresses the dynamic task assignment problem for multiple uncrewed aerial vehicles (UAVs) operating under weak communication. Existing learning-based methods face two primary challenges: limited scene generalization and excessive reliance on communication resources. To address these issues, this paper proposes an event-triggered reinforcement learning algorithm based on graph neural networks. First, heterogeneous UAVs and tasks are embedded into a graph to construct a relationship model, which clearly represents the complex constraints between UAVs and tasks. This graph structure overcomes the limitations of existing methods in capturing constraint relationships. Second, the incorporation of heterogeneous graph neural networks and adaptive attention mechanisms enables effective learning of changes in adjacent node information, allowing the model to capture complex constraint relationships and environmental dynamics. This approach also addresses the lack of sensitivity to environmental changes observed in existing methods. Lastly, a dynamic synchronization mechanism is employed to update task assignment statuses in real time, preventing task conflicts and ensuring efficient allocation. Experimental results demonstrate that this method strikes a better balance between task assignment quality and efficiency. It performs well in untrained scenarios and significantly reduces communication resource consumption. These results highlight its promising potential for application in weak communication environments. Note to Practitioners—This paper addresses the dynamic task assignment problem in multi-UAV systems under weak communication conditions. It proposes an event-triggered graph reinforcement learning algorithm combining graph neural networks and an adaptive attention mechanism. This framework models UAVs and tasks as nodes in a heterogeneous graph, capturing their relationships and constraints. The algorithm updates task assignments in real time using a dynamic synchronization mechanism, minimizing conflicts and ensuring optimal task completion despite intermittent communication. Its key advantage is balancing task quality and efficiency. Compared to traditional methods, it performs better in various scenarios, especially in untrained environments, and adapts to dynamic task demands. This approach is particularly useful in defense, logistics, agriculture, and surveillance. However, it has limitations in extreme cases of frequent UAV sabotage, and future work could optimize it for fully distributed environments to improve adaptability and reliability.
Funding Information
  • National Science and Major Project (2021ZD0112702)
  • National Natural Science Foundation of China (62373100, 62233003)
  • Big Data Computing Center of Southeast University

This publication has 40 references indexed in Scilit: