ABSTRACT Cognitive radio vehicular ad hoc networks (CR‐VANETs) demand secure and efficient routing to cope with high mobility, spectrum variability, and adversarial threats. To address these challenges, this paper introduces neural circle‐driven finite element fusion with eel electro‐dynamic optimization (NCir‐FeF‐E2dO). This integrated framework combines pre‐processing, feature extraction, intelligent routing, and encryption. Smooth‐Gauss histogram normalization (SGHNN) reduces noise and normalizes inputs, while the scale‐calibrated transformer (SCT) extracts multi‐scale features for robust representation. Routing predictions are achieved using a finite element neural fusion network (FE‐NFN) enhanced by circle‐driven optimization (CDO), and path selection is refined through eel electro‐dynamic optimization (E2dO). For secure communication, the hyperchaotic system–Fibonacci Q‐matrix encryption ensures confidentiality and resistance against statistical and differential attacks. Simulation results demonstrate significant performance gains: routing efficiency of 99.05%, packet delivery ratio of 97.22%, 30% lower end‐to‐end delay, 15% higher throughput, 10% better energy efficiency, 15% lower communication overhead, encryption speed of 142 Mbps, and system recovery time of 4.7 ms. Compared with OPBRP, OCSR, UGAVs‐MDVF, DTE‐RR, LSTM, UER, and RL‐IoT, the proposed method consistently outperforms across all evaluated metrics. These findings establish NCir‐FeF‐E2dO as a reliable and scalable solution for enhancing secure routing in CR‐VANETs.
Arunachalavel et al. (Fri,) studied this question.