Blockchain-Assisted Machine Learning For Securing Mobile Ad-Hoc Networks Against Black-Hole Attacks

Abstract

The Mobile Ad-Hoc Networks are dynamic and decentralized networks characterized by high mobility limited bandwidth and frequent topology changes This presents significant challenges for reliable and efficient routing. Traditional routing protocols face difficulties in adapting to network dynamics. This results in performance degradation and security vulnerabilities. This paper proposes a convolutional structure based on an adaptive neural fuzzy inference system, which is constantly evolving to maintain reliable communication in MANETs. The hybrid ANFIS model combines the knowledge capabilities of the structure. Neural networks match the decomposition power of fuzzy logic. This enables real-time intelligent rotation, decisions, dynamically adjusting to network conditions. Including moving nodes connection failures and congestion, ensuring robust route selection and early reliability management. Simulations performed on different MANET scenarios demonstrate the effectiveness of the structure in reducing packet loss. Improve shipping rates and enhance security by mitigating black hole and grey hole roaming attacks. The results indicate that the proposed model outperforms conventional protocols. and provides a scalable solution for future MANET deployments. This work contributes to building a more adaptable and secure network.