International Research Journal of Education and Technology

Mobile Ad Hoc Networks (MANETs) are susceptible to a number of security risks due to their high degree of dynamic nature and decentralized management. This work proposes a novel method to improve MANET security by fusing cryptography and artificial intelligence (AI) approaches. To be more precise, the combination of Random Forest classifiers and Long Short-Term Memory (LSTM) networks is suggested for anomaly detection, while cryptographic techniques such as AES and RSA are used to protect communications. Results show enhanced detection rates and secure routing in dynamic MANET environments, based on simulations using ns-3 and OMNeT++.