Maintaining Sensing Freshness: Task-Oriented Age of Information Based Joint Scheduling for UAV-Assisted ISCC Networks

With the rapid development of $\mathbf{6 G}$, the Integration of Sensing, Communication, and Computation (ISCC) has emerged as a critical architecture to support real-time, intelligent, and autonomous services in future smart cities and intelligent transportation systems. Unmanned Aerial Vehicles (UAVs), owing to their high flexibility and mobility, have become key enablers in ISCC networks. However, most existing studies on UAV-assisted networks mainly focus on partial components of sensing, communication, and computation, without fully considering their joint impact, leading to difficulties in ensuring end-to-end information freshness. To address this problem, building upon the concept of Age of Information (AoI), we specifically design Task-oriented AoI (TAoI). This metric enables a comprehensive evaluation of information timeliness across the entire processing pipeline, from raw sensing data to final results, by accounting for factors such as sensing update intervals, transmission failures, and computation delays. Based on the TAoI metric, we further develop a joint trajectory and task scheduling scheme to ensure the freshness of sensing data. To achieve collective intelligence in multi-UAV systems, we design a multi-agent deep reinforcement learning (MADRL) framework that enables autonomous and efficient decision-making for trajectory planning and task scheduling.