AdaptNet: Rethinking Sensing and Communication for a Seamless Internet of Drones Experience

TL;DR

The paper tackles safety, latency, and bandwidth challenges in dynamic IoD networks by integrating ISAC with $Fréchet distance$-based sensing and data-relevance thresholds. AdaptNet leverages two cooperative learning modes: MARL for sensing and MADDPG for communication, guided by a $Fréchet distance$-driven clustering and an $AoI$-based prioritization policy. Key contributions include a $Fréchet distance$-driven clustering mechanism, adaptive MU-MIMO waveform selection, and a dual-mode learning framework that improves URLLC performance while scaling to larger UAV fleets. Empirical results demonstrate up to ~45% improvements in clustering quality and notable speedups in learning convergence, confirming the approach's feasibility for turning drones into information orchestrators. The work lays a foundation for robust, efficient IoD operations and points to future enhancements with edge computing and real-world validations.

Abstract

In the evolving era of Unmanned Aerial Vehicles (UAVs), the emphasis has moved from mere data collection to strategically obtaining timely and relevant data within the Internet of Drones (IoDs) ecosystem. However, the unpredictable conditions in dynamic IoDs pose safety challenges for drones. Addressing this, our approach introduces a multi-UAV framework using spatial-temporal clustering and the Frechet distance for enhancing reliability. Seamlessly coupled with Integrated Sensing and Communication (ISAC), it enhances the precision and agility of UAV networks. Our Multi-Agent Reinforcement Learning (MARL) mechanism ensures UAVs adapt strategies through ongoing environmental interactions and enhancing intelligent sensing. This focus ensures operational safety and efficiency, considering data capture and transmission viability. By evaluating the relevance of the sensed information, we can communicate only the most crucial data variations beyond a set threshold and optimize bandwidth usage. Our methodology transforms the UAV domain, transitioning drones from data gatherers to adept information orchestrators, establishing a benchmark for efficiency and adaptability in modern aerial systems.

Figures (6)

• Figure 1: The broad spectrum of use cases for UAVs employing ISAC.
• Figure 2: Visual Overview of AdaptNet Architecture and its workflow from intelligent data acquisition in the IoD to dynamic cluster prediction utilizing the Fréchet distance, and to cooperative algorithm enhancements for improved UAV network efficiency and adaptability. It includes detailed segments depicting the coordination of state, action, and feedback mechanisms within MARL and MADDPG frameworks, aimed at prioritizing sensing and communication. Furthermore, it highlights AdaptNet's strength in prioritizing data relevance through adaptive waveform selection, showcasing its effectiveness across various network operations
• Figure 3: (a) Optimal sensing path of the UAVs along with the clustered fast-moving targets using Frechet distance; (b) Operational flow of Frechet distance analysis in the IoD framework. This block diagram shows the sequential process from initial target sensing to the final computation of Frechet distance. It starts with target detection and advances through preprocessing for trajectory input formulation. Subsequently, it undertakes geometric pattern analysis of these trajectories, engages in threshold comparisons for pattern similarity, and performs the precise calculation of the Frechet distance for effective target clustering and prioritization.
• Figure 4: (a) A comparison of UAV trajectories in ISAC analysis by revealing sensing anomalies in UAV1's path where the shaded region indicates a potential zone of instability; (b) The scaling impact of ISAC performance in UAV networks with increasing drone numbers.
• Figure 5: Comparison of AoI trends in IoD network with different queuing scenarios for ISAC.