Abstract:
In the evolving Internet of Vehicles (IoV) landscape, ensuring security and
efficiency is critical. This study integrates Federated Learning (FL) with edge
caching to address these challenges. FL allows decentralized model training,
safeguarding data privacy and reducing communication overhead. Coupled with
edge caching, latency and bandwidth consumption are minimized by storing
frequently accessed data at the network edge. Our approach optimizes resource
use with parameter tuning, momentum for faster convergence, and parallel
processing for scalability. Results demonstrate that this combination
significantly enhances IoV network performance, paving the way for secure,
efficient vehicular communication systems
