UAV-Enabled Wireless Networks for Integrated Sensing and Learning-Oriented Communication

TL;DR

This paper proposes a UAV-enabled wireless network that jointly optimizes sensing and learning services, introducing a novel algorithm to enhance AI learning performance while maintaining sensing quality in integrated networks.

Contribution

It introduces a convergence-guaranteed iterative algorithm for joint optimization of UAV trajectory, power, and time allocation in sensing and learning tasks.

Findings

The proposed algorithm outperforms baseline methods in simulations.

There is a critical tradeoff between sensing quality and learning performance.

Joint optimization significantly improves overall network efficiency.

Abstract

Future wireless networks are envisioned to support both sensing and artificial intelligence (AI) services. However, conventional integrated sensing and communication (ISAC) networks may not be suitable due to the ignorance of diverse task-specific data utilities in different AI applications. In this letter, a full-duplex unmanned aerial vehicle (UAV)-enabled wireless network providing sensing and edge learning services is investigated. To maximize the learning performance while ensuring sensing quality, a convergence-guaranteed iterative algorithm is developed to jointly determine the uplink time allocation, as well as UAV trajectory and transmit power. Simulation results show that the proposed algorithm significantly outperforms the baselines and demonstrate the critical tradeoff between sensing and learning performance.