A Survey of Methods and Architectures for Dynamic Path-Controllable Deep Unfolding Network to predict the K-barriers for intrusion detection using a wireless sensor network

Abstract

Wireless Sensor Networks (WSNs) play a vital role in surveillance and intrusion detection applications, where reliable monitoring of unauthorized movement is essential. An important concept in this field is k barrier coverage, which ensures that every intrusion path intersects at least k sensor barriers, thereby improving detection reliability and fault tolerance. However, dynamic k barrier prediction remains challenging because of node mobility, limited resources, communication constraints, and real time operational demands. Recent advances in Dynamic Path Controllable Deep Unfolding Networks (DPDUNs) have provided promising solutions by integrating optimization based methods with deep learning techniques. These hybrid architectures transform iterative optimization processes into neural network layers, enabling adaptive learning in dynamic WSN environments. Combined with spatial temporal models, reinforcement learning, and attention mechanisms, DPDUNs improve barrier prediction accuracy and optimize intrusion detection paths under changing network conditions. This survey reviews 30 studies published between 2020 and 2023 on deep unfolding networks, hybrid AI models, temporal learning, optimization strategies, and AI driven intrusion detection frameworks for WSNs. The review highlights major trends, strengths, limitations, and research gaps. Although hybrid AI approaches show superior performance, challenges such as scalability, deployment complexity, energy efficiency, and real time adaptability still require further research.