Enhanced Quantum-Inspired Metaheuristic Framework for Federated Trust Management and Blockchain-Enabled Privacy in Large-Scale Internet of Vehicles

The dynamic nature and large scale of the Internet of Vehicles (IoV) pose challenges in managing scalable, trust, and privacy-preserving trust management systems. The conventional centralized approaches face vulnerability issues in privacy breaches and single-point failures, whereas traditional federated learning techniques are complex to explore non-IID data distribution and exhibit low convergence in intricate IoV topologies. These limitations are overcome by introducing the Enhanced Quantum-Inspired Metaheuristic Framework for Federated Learning (EQIMFFL) to ensure a secure and trusted IoV communication environment. Initially, quantum-inspired networks are applied to a trust model that utilizes quantum-inspired activation and superposition-related neurons to enhance feature exploration for trust inference. A particle optimization-related federated aggregation process is incorporated to train the vehicle features and update them to the roadside units and central servers. The hybrid optimization process implemented at the RSU level enhances the aggregation, which updates the model and directly improves the global model accuracy. Finally, a blockchain ledger is incorporated to maintain the cloud server and RSU, in which the hash function is integrated to manage records that directly impact trust decisions, confirming the resilience against unwanted activities, auditability, and integrity. In addition, the Shapley value is incorporated to reward the vehicle for maintaining the data quality while ensuring secure transmission. Then, the complex urban IoV scenario is implemented to evaluate the system’s performance, in which the framework achieves a 33.1% improvement in communication overhead, 18.5% faster convergence, and 96.83% trust prediction accuracy compared to the baseline methods.