2026 ASIAN Conference on Communication and Networks (ASIANComNet 2026)

Unmanned aerial vehicles (UAVs) are expected to play an important role in sixth-generation (6G) integrated sensing and communication (ISAC) networks, shaping the future world by providing flexible connectivity and real-time situational awareness. However, coordinating multiple UAVs in dynamic environments requires the joint optimization of mobility, sensing, communication resource allocation, energy consumption, and operational safety. This work proposes a multi-agent deep reinforcement learning (MADRL) framework for cooperative UAV operations in 6G ISAC networks. Each UAV acts as an autonomous agent that jointly determines its movement and ISAC resource-allocation mode based on local observations. A multi-agent proximal policy optimization approach with centralized training and decentralized execution is employed to learn cooperative policies while enabling scalable local decision-making during deployment. The reward function jointly accounts for target detection, communication quality, information acquisition, collision avoidance, energy consumption, and stable ISAC allocation. Simulation results show that the proposed framework achieves a higher average system reward and shorter target-detection time than rule-based and single-agent DRL schemes, while maintaining a favorable trade-off between mission performance and energy consumption. These results demonstrate the potential of cooperative MADRL to support intelligent, adaptive, and reliable UAV operations in future 6G ISAC networks. 

Unmanned aerial vehicle (UAV),6G ISAC,Multi-Agent DRL,Autonomous Systems