Semantic-Aware Edge Intelligence for Cooperative Perception in UAV–V2X Mobility
Abstract
This chapter presents a semantic-aware edge intelligence framework for cooperative perception in UAV–V2X mobility. It addresses the limitations of continuous raw-data sharing among vehicles, UAVs, roadside units, MEC servers, and cloud platforms under bandwidth, latency, energy, synchronization, and computing constraints. The framework shifts cooperation from data-heavy exchange to task-oriented sharing of meaningful information, including occlusion alerts, uncertainty indicators, risk levels, motion intentions, and compact feature representations. It integrates semantic relevance assessment, hierarchical edge execution, split computing, federated learning, and learning-assisted orchestration into a unified seven-layer architecture. The chapter also provides lightweight formulations, practical design recommendations, deployment scenarios, and future research directions for scalable, reliable, privacy-aware, and 6G-ready cooperative perception systems.
