Artificial Intelligence in Smart Waste Management: A Systematic Review of Algorithms, Integrated Architectures, and Future Research Directions

Abstract

Urbanization, industrial development and changing consumption patterns have also been factors in ongoing growth of municipal solid waste (MSW) generation, which places extreme pressure on urban infrastructure, environmental quality and public health. Yet, traditional waste management approaches that focus predominantly on fixed collection times and manual sorting face challenges in contexts where the volume of production is fluid – resulting in overflowing bins, poor segregation of waste streams, ineffective routing of vehicles and higher emissions. These limitations have been addressed by smart waste management systems which have shifted from IoT monitoring (fill-level sensing) to AI systems that can be predictive, adaptive, and self-optimizing. This paper summarizes the available literature on various Artificial Intelligence (AI) technologies used for smart waste management, with a particular focus on detection/classification, fill level estimation, monitoring and operational decision-making. To be transparent and reproducible, a PRISMA-guided screening process was used: 780 records were identified, 602 records after duplicates, 450 records were screened, 150 were full texts and after final synthesis 128 studies were included. Results indicate research is typically spread out across individual components, and there is a lack of standardization in datasets, benchmarking and widespread implementation of real-world validation, which calls for an integrated architecture. Therefore, this study suggests a YOLOv8-based smart waste framework that integrates real-time waste detection/classification, waste bin fill-level estimation (empty/half-full/full) and monitoring/decision layer, which can be delivered at edge or hybrid (edge–cloud) architecture for low latency operation and scalability. The direction of the proposed concept is towards designing a unified system to increase the efficiency of the operations, decrease manual efforts, and increase sustainable, data-driven waste management in cities.