Federated Distillation Assisted Vehicle Edge Caching Scheme Based on Lightweight DDPM

TL;DR

The paper addresses privacy-preserving, low-communication vehicle edge caching in highly mobile environments. It introduces a federated distillation framework built on a lightweight DDPM (LDPM), where vehicles share hash-encoded data and knowledge instead of raw data, and RSUs fuse neighbor knowledge via cosine similarity to personalize models. Local LDPMs trained with aggregated knowledge predict content popularity, while mobility-aware cache replacement and KC synchronization optimize cache content across RSUs and the MBS. Experimental results on MovieLens 1M demonstrate improved cache hit rates and substantially reduced communication overhead, highlighting practical benefits for real-world vehicular networks.

Abstract

Vehicle edge caching is a promising technology that can significantly reduce the latency for vehicle users (VUs) to access content by pre-caching user-interested content at edge nodes. It is crucial to accurately predict the content that VUs are interested in without exposing their privacy. Traditional federated learning (FL) can protect user privacy by sharing models rather than raw data. However, the training of FL requires frequent model transmission, which can result in significant communication overhead. Additionally, vehicles may leave the road side unit (RSU) coverage area before training is completed, leading to training failures. To address these issues, in this letter, we propose a federated distillation-assisted vehicle edge caching scheme based on lightweight denoising diffusion probabilistic model (LDPM). The simulation results demonstrate that the proposed vehicle edge caching scheme has good robustness to variations in vehicle speed, significantly reducing communication overhead and improving cache hit percentage.

Figures (5)

• Figure 1: System Model.
• Figure 2: Execution between Vehicles and RSUs.
• Figure 3: Loss versus episodes.
• Figure 4: Request content delay versus cache capacity.
• Figure 5: Cache hit percentage versus vehicle speed.