A Predictive UAV Framework for Tracking Fast-Moving Vehicles in Dynamic Environments

Authors:
Publisher website
Google Scholar
Abstract
In the near future, Uncrewed Aerial Vehicles (UAVs) are expected to evolve from simple data collectors to intelligent data gatekeepers capable of processing and analyzing information directly on board. In this paper, a decentralized UAV framework is proposed to address potential challenges in communication and coordination among fast-moving vehicles due to difficulties in maintaining seamless information exchange and synchronized operation across the network. While Integrated Sensing and Communications (ISAC) provides significant advantages, it still faces considerable challenges when dealing with such scenarios. This paper proposes DynaMo, a decentralized UAV framework addressing the challenges of tracking fast-moving vehicles in dynamic environments. DynaMo prioritizes spatiotemporal relevance and adapts to diverse vehicle behaviors with a novel freshness metric. Integrated with a Partially Observable Markov Decision Process (POMDP), the framework enables UAVs to make informed, real-time decisions under uncertainty. Simulation results highlight the superior tracking precision and adaptability of our proposed approach, offering a robust solution for ISAC systems in high-mobility scenarios.

This publication has 22 references indexed in Scilit: