Hybrid Deep Learning for Secure IoT-WSN Environments and Renewable Energy Forecasting Using Tunicate and Rooster Optimization

Abstract

In the contemporary era of interconnected intelligent systems, the Internet of Things (IoT) and Wireless Sensor Networks (WSNs) form the backbone of real-time data acquisition, automation, and smart energy management. However, the vast interconnectivity among devices exposes these networks to multifaceted cyber threats and energy inefficiencies, demanding robust, intelligent, and adaptive mechanisms for security and forecasting. This paper introduces a hybrid deep learning framework that unifies Tunicate Swarm Optimization (TSO) and Rooster Optimization Algorithm (ROA) to enhance both cybersecurity in IoT-WSN systems and renewable energy forecasting accuracy. The TSO algorithm improves feature selection and dimensionality reduction, while ROA optimizes the deep neural architecture parameters, leading to superior detection accuracy and energy efficiency. The hybrid model integrates Deep Belief Networks (DBN) and Long Short-Term Memory (LSTM) networks to simultaneously handle temporal dependencies and feature complexities in IoT data streams. The experimental analysis demonstrates that the proposed model achieves significant improvements in intrusion detection rates, energy consumption prediction, and overall computational efficiency compared to existing benchmark techniques. This study provides a comprehensive framework for sustainable, secure, and intelligent IoT-WSN environments, bridging the gap between cybersecurity and renewable energy forecasting within a unified optimization paradigm.