An Efficient E-Waste Management System through Energy-Aware Routing and Hybrid Optimization Deep Learning Routing on an IoT-Cloud Platform

The rapid pace of technological expansion has resulted in an increase in electronic waste (e-waste) generation. The management of e-waste is a complex and labour-intensive process that requires significant resources. To address this issue, research has focused on developing advanced technologies to classify and manage e-waste more efficiently. In this study, a hybrid approach combining Fractional Henry Gas Optimization (FHGO) and Horse Herd Optimization Algorithm (HOA) is proposed to classify e-waste. The proposed approach, named Fractional Horse Herd Gas Optimization-based Shepherd Convolutional Neural Network (FrHHGO-based ShCNN), is implemented on an IUoT-cloud platform-waste images are collected by IoT nodes and stored in cloud storage, which then uses the FrHHGO algorithm for routing. Feature extraction procedures, such as the Gray Level Co-occurrence Matrix (GLCM) feature, Local Gabor Binary Pattern (LGBP), and Histogram of Oriented Gradient (HOG), are performed, and data augmentation is used to supplement the extracted feature size. The proposed approach outperformed several existing approaches such as Deep Learning, Tensor flow deep learning, Cuckoo search-based Neural network, and Machine learning, achieving an accuracy of 19.49%, 18.05%, 12.77%, and 7.89%, respectively. Moreover, the developed method achieved better performance metrics, including minimal energy of 0.301 J, minimal delay of 0.666 sec, maximal accuracy of 0.950, maximal sensitivity of 0.934, and maximal specificity of 0.967. Thus, the developed FrHHGO-based ShCNN e-waste classification system enhances the social, environmental, and economic sustainability of emerging economies.