Edge Computing Resource Optimization for UAV Collaboration in Sensing, Computing and Communication Networks

This chapter delves into the optimization of resource allocation in UAV-ISCC networks, focusing on minimizing energy consumption while ensuring robust sensing performance. The study introduces a multi-agent federated deep reinforcement learning (MAFD) algorithm to tackle the complexities of multi-agent collaboration in dynamic environments. The text explores the system model, including sensing, communication, and computing models, and formulates the optimization problem with constraints on task offload rate, sensing decision, task delay, transmission power, and perceptual performance. The proposed MAFD algorithm is compared with other algorithms like DQN and DDPG, demonstrating superior performance in terms of reward and energy consumption. Simulation results highlight the effectiveness of the MAFD algorithm, showing a significant reduction in energy consumption. The conclusion emphasizes the potential of the MAFD algorithm for optimizing system performance in UAV-ISCC networks and suggests areas for future research, such as dynamic scenarios and UAV flight time considerations.