Co-operative deep belief fused deep neuro-fuzzy network for water depth-aware content caching model in underwater IoT

Abstract

The Internet of Underwater Things (IoUT) faces challenges in effective storage and large-scale content distribution due to its reliance on Internet Protocol (IP), which is not optimized for underwater environments. Information Centric Networking (ICN) offers a promising solution to address these concerns, but its performance is hindered by issues such as propagation speed, delay, and latency, making traditional ICN models unsuitable for IoUT. To overcome these limitations, this work proposes a novel water-depth aware content caching model, the Co-operative Deep Belief Network fused Deep Neuro-Fuzzy Network (DBNFDNFN), which is formed by the integration of Deep Belief Networks (DBN) and Deep Neuro-Fuzzy Networks (DNFN). This model introduces three main modules: the DBN module, the Caching Content Layer, and the DNFN module. Initially, the IoUT nodes are simulated within an underwater sensor network, where naming enables users to express required information and sensors to accurately describe their data. The communication mode ensures efficient packet forwarding and retrieval. Finally, the Co-operative DBNFDNFN model optimizes content caching decisions. Moreover, the proposed Co-Operative DBNFDNFN model achieves a minimal delay of 0.0067 s, low power consumption of 11.223W, a high cache hit rate of 41, and a fast-running time of 1.237 s, highlighting its effectiveness and efficiency for IoUT applications. Additionally, in this work, the DBN processes large datasets to predict future requests, while the DNFN uses fuzzy logic for flexible, adaptive decision-making. Together, they enable dynamic caching adjustments, ensuring optimal performance in varying network and environmental conditions.