FedAPT: Joint Adaptive Parameter Freezing and Resource Allocation for Communication-Efficient Federated Vehicular Networks

IEEE Internet of Things Journal, Journal Year: 2024, Volume and Issue: 11(11), P. 19520 - 19536

Published: March 13, 2024

Telematics technology development offers vehicles a range of intelligent and convenient functions, including navigation mapping services, driving assistance, traffic management. However, since these functions deal with sensitive information like vehicle location habits, it is crucial to address concerns regarding security privacy protection. Federated learning (FL) highly suitable for addressing such problems due its characteristics, in which client does not need share private data upload model parameters parameter server via the network. This results establishment federated network (FVN). As distributed paradigm, efficiency communication as impacts all aspects FVN. paper introduces freezing algorithm based on historical reduce transferred between each round communication, thus minimizing overhead learning. Additionally, we propose using particle swarm allocate bandwidth packet sizes sent by (i.e., non-freezing parameters) minimize latency FL round. Furthermore, high time complexity algorithm, employ generate training transformer fast response sufficient accuracy, thereby accelerating allocation process. Through extensive experiments, prove feasibility our approach improving

Language: Английский

Contract-based resource reservation for energy harvesting-enabled mobile edge computing

The Journal of Supercomputing, Journal Year: 2025, Volume and Issue: 81(5)

Published: April 3, 2025

Language: Английский

Opportunistic Routing for Mobile Edge Computing: A Community Detected and Task Priority Aware Approach

Published: Jan. 1, 2024

Language: Английский