Improving Manet Security using Crayfish Optimization and GRU-LSTM Classifier for Intrusion Detection and Prevention

Mobile Ad-Hoc Networks (MANETs) face significant security challenges owing to their dynamic nature, distributed structure and frequent mobility of the nodes. An efficient Intrusion Detection and Prevention Systems (IDPS) is necessary to provide security against malicious attacks or disruptions to its operation. Although traditional methods utilize a selection of Machine Learning (ML) and Deep Learning (DL) methods, susceptible to increased computational complexity, decreased system performance and high false-positive rates. Therefore, this paper presents a hybrid framework to protect MANETs by combining Crayfish Optimization Algorithm (COA) with a Gated Recurrent Units (GRU) and Long Short-Term Memory (LSTM) classifier for Intrusion Detection (ID) and prevention. COA used for enhancing MANET routing by identifying the shortest and efficient path for data transmission. The hybrid model classify/predict different types of intrusions accurately by integrating the complementary strengths of GRU-LSTM models. The proposed approach is implemented using Python software with WSN-DS datasets through various experiments demonstrating significant improvements in ID rates, accuracy of 95%, false positive rates, and system efficiencies to provide a robust approach to securing MANETs.