MANET’s Node Secure Mobility Predictions using Enhanced Adaptive Learning Techniques

Predicting node mobility in dynamic environments is crucial to designing and implementing ad hoc networks. This study presents a filter-based computing approach. Each node's mobility prediction model predicts its neighbors' movements. It uses node spatial and temporal properties. This research suggests using reinforcement learning to increase model accuracy. The greeting message threshold determines the best neighbor node-finding strategy. HP-AODV and ROMSG are used to assess the paper's performance. The proposed welcome message broadcasting algorithm is far cheaper than others. It reduces neighbor node discovery errors. This strategy improves ad hoc wireless network quality. Mobility prediction model integration into the network layer is complex. However, application-level integration may improve routing protocol efficiency. The study develops a mobility prediction framework to anticipate a wireless device's future position reliably. The mobility prediction model is a sequence of discrete occurrences, such as a node's upcoming position, based on its present location. The research suggests using the AdaBoost algorithm and Markov model to increase accuracy. AdaBoost estimates model weight coefficients. The AdaBoost-produced multi-order Markov model beats conventional Markov models.