Two-Layer Federated Learning With Heterogeneous Model Aggregation for 6G Supported Internet of Vehicles

The vision of the upcoming 6G technologies that have fast data rate, low latency, and ultra-dense network, draws great attentions to Internet Vehicles (IoV) Vehicle-to-Everything (V2X) communication for intelligent transportation systems. There is an urgent need distributed machine learning techniques can take advantages massive interconnected networks with explosive amount heterogeneous generated at network edge. In this study, a two-layer federated model proposed end-edge-cloud architecture typical in environment, achieve more efficient accurate while ensuring privacy protection reducing overheads. A novel multi-layer selection aggregation scheme designed as part process better utilize local global contexts individual vehicles road side units (RSUs) supported vehicular networks. This context-aware mechanism then developed applied address object detection, which one most critical challenges modern systems autonomous vehicles. Evaluation results showed method, demonstrates higher accuracy precision, recall F1 score, outperforms other state-of-the-art methods under configuration by achieving faster convergence, scales larger numbers RSUs involved process.