Federated Trust-Aware Spectrum Sensing For Cognitive Radio Enabled Internet Of Vehicles In 6G Networks

Abstract: The rapid evolution of sixth-generation (6G) wireless networks has accelerated the development of intelligent communications systems capable of supporting ultra-reliable, low-latency applications.Cognitive Radio-Enabled Internet of Vehicles (CR-IoV) has emerged as a promising solution for efficient spectrum utilization in highly dynamic automotive environment. However, conventional centralized spectrum sensing techniques suffer from high communication overhead, privacy issues, and scalability limitations when deployed in large-scale vehicular networks. To address these challenges, this paper proposes a Federated Trust-Aware Spectrum Sensing framework for CR-IoV networks operating in 6G environments. The proposed approach integrates Federated Learning (FL) with a trust evaluation mechanism to enable decentralized model training while preserving user privacy and improving detection reliabilit. Convolutional Neural Networks (CNN) are used for local spectrum occupancy detection, while support vector machines (SVM) are used for intelligent road unit (RSU) selection and resource allocation. Trust-based aggregation is integrated to mitigate the impact of untrusted and malicious nodes during the global model updating process. MATLAB simulations are performed under different signal-to-noise ratio (SNR) conditions to evaluate the performance of the proposed framework. Simulation results demonstrate significant improvements in detection probability, detection accuracy, throughput, and spectrum utilization compared to centralized and conventional federated learning approaches. The proposed framework provides a scalable, secure and efficient solution for future 6G-enabled intelligent vehicle communication systems.

Keywords: Cognitive Radio, Internet of Vehicles, Federated Learning, Trust Management, Spectrum Sensing, Convolutional Neural Network, Support Vector Machine, 6G Networks.