Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimizing NOMA Transmissions to Advance Federated Learning in Vehicular Networks

Ziru Chen, Zhou Ni, Peiyuan Guan, Lu Wang, Lin X. Cai, Morteza Hashemi, Zongzhi Li

TL;DR

This work tackles privacy-preserving learning in vehicular networks by employing Federated Learning across vehicles, where data never leaves local devices. It introduces a NOMA-enabled Federated Vehicular Network (FVN) framework that jointly optimizes vehicle selection and uplink power allocation to maximize the joining ratio $M_t$ under SINR constraints, using successive interference cancellation at the base station. An efficient heuristic (Algorithm 1) for vehicle selection and power control is paired with an NFL extension of FedAvg (Algorithm 2) to enhance convergence and stability across i.i.d. and non-i.i.d. data, demonstrated on CIFAR-10 with 80 clients. Results show the proposed approach outperforms OMA and CVX-based baselines in both joining ratio and learning performance, indicating significant practical potential for scalable, privacy-preserving FL in dynamic vehicular networks.

Abstract

Diverse critical data, such as location information and driving patterns, can be collected by IoT devices in vehicular networks to improve driving experiences and road safety. However, drivers are often reluctant to share their data due to privacy concerns. The Federated Vehicular Network (FVN) is a promising technology that tackles these concerns by transmitting model parameters instead of raw data, thereby protecting the privacy of drivers. Nevertheless, the performance of Federated Learning (FL) in a vehicular network depends on the joining ratio, which is restricted by the limited available wireless resources. To address these challenges, this paper proposes to apply Non-Orthogonal Multiple Access (NOMA) to improve the joining ratio in a FVN. Specifically, a vehicle selection and transmission power control algorithm is developed to exploit the power domain differences in the received signal to ensure the maximum number of vehicles capable of joining the FVN. Our simulation results demonstrate that the proposed NOMA-based strategy increases the joining ratio and significantly enhances the performance of the FVN.

Optimizing NOMA Transmissions to Advance Federated Learning in Vehicular Networks

TL;DR

This work tackles privacy-preserving learning in vehicular networks by employing Federated Learning across vehicles, where data never leaves local devices. It introduces a NOMA-enabled Federated Vehicular Network (FVN) framework that jointly optimizes vehicle selection and uplink power allocation to maximize the joining ratio under SINR constraints, using successive interference cancellation at the base station. An efficient heuristic (Algorithm 1) for vehicle selection and power control is paired with an NFL extension of FedAvg (Algorithm 2) to enhance convergence and stability across i.i.d. and non-i.i.d. data, demonstrated on CIFAR-10 with 80 clients. Results show the proposed approach outperforms OMA and CVX-based baselines in both joining ratio and learning performance, indicating significant practical potential for scalable, privacy-preserving FL in dynamic vehicular networks.

Abstract

Diverse critical data, such as location information and driving patterns, can be collected by IoT devices in vehicular networks to improve driving experiences and road safety. However, drivers are often reluctant to share their data due to privacy concerns. The Federated Vehicular Network (FVN) is a promising technology that tackles these concerns by transmitting model parameters instead of raw data, thereby protecting the privacy of drivers. Nevertheless, the performance of Federated Learning (FL) in a vehicular network depends on the joining ratio, which is restricted by the limited available wireless resources. To address these challenges, this paper proposes to apply Non-Orthogonal Multiple Access (NOMA) to improve the joining ratio in a FVN. Specifically, a vehicle selection and transmission power control algorithm is developed to exploit the power domain differences in the received signal to ensure the maximum number of vehicles capable of joining the FVN. Our simulation results demonstrate that the proposed NOMA-based strategy increases the joining ratio and significantly enhances the performance of the FVN.
Paper Structure (8 sections, 3 theorems, 14 equations, 4 figures, 2 algorithms)

This paper contains 8 sections, 3 theorems, 14 equations, 4 figures, 2 algorithms.

Table of Contents

  1. Introduction
  2. System Model
  3. NOMA-enabled FVN framework
  4. Data Communication Model
  5. Vehicle Selection and Transmission Power Allocation
  6. Federated Learning
  7. Performance Evaluation
  8. Conclusion

Key Result

Lemma 1

Provided that the signal from the $(n+1)$-th vehicle is decodable, the signal from the $n$-th vehicle will be decodable if the following conditions are met: $\frac{p^r_{n,t} }{p^r_{n+1,t}}\geq (1+\Gamma), \forall n<M_t-1$; and $p^r_{M_t-1,t}\geq (p^r_{M_t,t}+\sigma_t^2)\Gamma$.

Figures (4)

  • Figure 1: System Model Illustration.
  • Figure 2: Average number of successfully connected vehicles versus the total number of vehicles in the network.
  • Figure 3: Average number of successfully connected vehicles versus the total number of FL joining ratio.
  • Figure 4: Federated Learning results with CIFAR-10 data.

Theorems & Definitions (6)

  • Lemma 1
  • proof
  • Lemma 2
  • proof
  • Lemma 3
  • proof