Clustered Vehicular Federated Learning: Process and Optimization

Federated Learning (FL) is expected to play a prominent role for privacy-preserving machine learning (ML) in autonomous vehicles. FL involves the collaborative training of single ML model among edge devices on their distributed datasets while keeping data locally. While requires less communication compared classical learning, it remains hard scale large models. In vehicular networks, must be adapted limited resources, mobility nodes, and statistical heterogeneity distributions. Indeed, judicious utilization resources alongside new perceptive learning-oriented methods are vital. To this end, we propose architecture corresponding scheduling processes. The utilizes vehicular-to-vehicular(V2V) bypass bottleneck where clusters vehicles train models simultaneously only aggregate each cluster sent multi-access (MEC) server. formation multi-task takes into account both aspects. We show through simulations that proposed process capable improving accuracy several non-independent and-identically-distributed (non-i.i.d) unbalanced distributions, under constraints, comparison standard FL.