A Comprehensive Review of Prediction of Scenarios for Routing in IoT-Based MANETs Using Expanding Ring Search and Random Early Detection Parameters with Global Pooling Dilated Convolutional Neural Networks

Abstract

Mobile Ad Hoc Networks (MANETs) integrated with Internet of Things (IoT) devices create highly dynamic communication environments where efficient routing is critical for reliable performance. Traditional routing protocols often struggle with issues such as frequent topology changes, node mobility, congestion, and limited resources. Techniques like Expanding Ring Search (ERS) and Random Early Detection (RED) have been widely used to improve routing efficiency and congestion control, where ERS minimizes routing overhead and RED prevents congestion through proactive packet management. Recent advancements incorporate machine learning and deep learning techniques to predict network conditions and optimize routing parameters. By analyzing ERS and RED-related metrics, these models can estimate throughput, packet delivery ratio, and delay, enabling adaptive Quality of Service (QoS). In particular, Global Pooling Dilated Convolutional Neural Networks enhance scenario prediction by capturing broader contextual information while maintaining computational efficiency. This review highlights the integration of ERS, RED, and deep learning models, discussing current methodologies, challenges, and future directions for intelligent routing optimization in IoT-based MANET systems.