On Comprehensive Review and Simulation Analyses of Intelligent Autonomous Waste Management Systems

Growth in population and consumerism has triggered an unprecedented increase in waste generation. However, the existing waste management systems are outstripped and do not promote sustainable living. An intelligent autonomous waste management system, discussed in this paper, addresses these challenges by integrating swarm intelligence and machine learning. Various object detection models and convolutional neural network (CNN) architectures, including ResNet, and their fusions are analysed for waste classification accuracy. IoT-based waste bin level and gas monitoring are explored for effective waste tracking. Various swarm intelligence algorithms, including particle swarm optimization (PSO), are reviewed for optimized waste collection. Image classification models such as You Only Look Once (YOLO) and Faster R-CNN are evaluated for accuracy and speed in waste classification. From the simulation results, it is demonstrated that YOLOv8 is the best-suited real-time waste classification model, achieving an accuracy of 95.8%. Further, the proposed PSO algorithm is assessed for effective waste collection routes by reducing environmental impact. From the simulation results obtained through the proposed PSO-based waste collection algorithm, it has been proven to outperform standard PSO. This study demonstrates the efficacy of integrating swarm intelligence, machine learning, and IoT in waste management and their potential to transform the waste management process.