ProWaste for proactive urban waste management using IoT and machine learning – Nature

 

Executive Summary: The ProWaste Initiative for Sustainable Urban Development

Reactive waste management in rapidly growing urban centers presents a significant challenge to achieving the Sustainable Development Goals (SDGs). Overflowing Waste-Collection Centres (WCCs) directly threaten SDG 3 (Good Health and Well-being) through odour and leachate, and undermine SDG 11 (Sustainable Cities and Communities) by degrading the urban environment. This report details ProWaste, an end-to-end Internet-of-Things (IoT) and machine-learning platform designed to transition municipal waste collection from a reactive to a proactive, data-driven model. By integrating fifteen automated and manual indicators, the system predicts WCC service needs with over 99% accuracy using just three key features. This innovation supports SDG 9 (Industry, Innovation, and Infrastructure) by creating resilient and efficient infrastructure. The ProWaste platform provides a scalable and transferable solution that eliminates missed pickups, reduces unnecessary vehicle inspections, and optimizes resource use, directly contributing to SDG 12 (Responsible Consumption and Production) and SDG 13 (Climate Action).

1. Introduction: Urban Waste Management as a Critical SDG Challenge

The rapid expansion of urban populations has intensified municipal solid waste generation, posing a critical threat to sustainable urban living. Current waste management practices are often inefficient and reactive, leading to overflowing collection centers. This systemic failure directly contravenes several Sustainable Development Goals:

• SDG 11 (Sustainable Cities and Communities): Inefficient waste collection degrades urban environments, reduces quality of life, and places unsustainable pressure on municipal services.
• SDG 3 (Good Health and Well-being): Waste overspills create public health hazards, including toxic fumes, pollutants, and the spread of disease vectors, affecting the well-being of residents, particularly in densely populated areas.
• SDG 12 (Responsible Consumption and Production): The lack of data-driven collection prevents the optimization of waste streams, hindering efforts towards recycling and a circular economy.

The ProWaste project was conceived to address these challenges by leveraging technology to create an intelligent, proactive waste management system. By forecasting the service needs of WCCs, the system aims to build more resilient, clean, and sustainable cities in line with the 2030 Agenda for Sustainable Development.

2. The ProWaste Framework: A System for Sustainable Impact

2.1. System Architecture and Alignment with SDGs

The ProWaste framework is an integrated system designed for proactive waste management, built on IoT and machine learning. Its components are strategically chosen to maximize sustainability and operational efficiency.

1. Data Acquisition: A network of low-cost sensors and public APIs gathers real-time data on fifteen indicators, including population density, weather, maintenance history, and waste accumulation. This promotes SDG 9 by utilizing affordable and innovative technology for infrastructure management.
2. Cloud Integration: Data is streamed to a Google Firebase cloud database, ensuring secure storage and accessibility for analysis and monitoring.
3. Machine Learning Module: A predictive model analyzes the integrated data to generate a “criticality score” for each WCC, forecasting which centers require immediate attention.
4. Sustainable Smart Waste Management (SSWM) Mobile Application: This mobile interface delivers real-time alerts and a dynamically prioritized maintenance queue to field teams. This empowers municipal workers and optimizes logistics, contributing to SDG 8 (Decent Work and Economic Growth) and SDG 13 (Climate Action) by reducing unnecessary vehicle mileage and associated emissions.

2.2. Input Parameters for Holistic Analysis

The model’s strength lies in its use of fifteen diverse, heterogeneous variables to create a comprehensive picture of waste generation dynamics. These inputs move beyond simple fill-level sensing to include contextual factors crucial for accurate prediction and sustainable planning.

• Demographic and Geographic Data: Population density, proximity to public centers, transit hubs, and water bodies.
• Environmental Data: Real-time temperature, humidity, precipitation, and air pollution scores (AQI), which influence decomposition rates and public health risks (SDG 3).
• Operational Data: Time since last maintenance and historical weekly waste aggregation data.
• Infrastructural and Social Data: Average housing structure, frequency of public gatherings, and condition of public infrastructure.

3. Methodology: Optimizing for Efficiency and Interpretability

3.1. Model Selection and Benchmarking

To ensure the selection of a robust and practical model, twenty-five different machine learning classifiers were benchmarked. While ensemble methods like AdaBoostClassifier showed the highest accuracy, a Decision Tree Classifier was chosen as the base learner. This decision was driven by the need for a model that balances high accuracy with interpretability and computational efficiency—key requirements for deployment by municipal authorities who need to understand and trust the system’s outputs (SDG 11.a, supporting positive links between urban and rural areas).

3.2. Feature Selection using Binary Particle Swarm Optimisation (BPSO)

A primary goal was to create a system that is not only accurate but also resource-efficient. A wrapper-based feature selection method using Binary Particle Swarm Optimisation (BPSO) was employed to identify the most impactful variables. The BPSO process successfully reduced the number of required inputs by 80%, from fifteen to just three, without compromising predictive power. This optimization is critical for scalability and affordability, aligning with SDG 9 and SDG 12 by minimizing data bandwidth and computational overhead.

4. Results: High-Accuracy Prediction with a Minimalist Model

4.1. Predictive Performance

The final, optimized model demonstrated exceptional performance on a hold-out test set, achieving over 99.8% accuracy, balanced accuracy, and macro-F1 score. The model misclassified only one record out of 1,391 instances. This high level of accuracy ensures that maintenance teams can rely on the system to prevent overspills, directly supporting the goals of cleaner and healthier cities (SDG 3, SDG 11).

4.2. Identification of Key Drivers of Waste Criticality

The BPSO feature selection process revealed that WCC criticality can be predicted with near-perfect accuracy using only three features:

1. Weekly Waste Accumulation
2. Time Since Last Maintenance (days)
3. Air Pollution (AQI)

The dominance of these three factors simplifies the data collection requirements for future deployments and provides clear, actionable insights for waste management authorities. SHAP (SHapley Additive exPlanations) analysis confirmed the interpretability and reliability of this three-feature model, showing that these variables consistently drive predictions across all criticality classes.

5. Conclusion: Advancing Sustainable Urban Futures with ProWaste

The ProWaste platform represents a significant step forward in applying technology to solve fundamental urban challenges in alignment with the Sustainable Development Goals. By shifting waste management from a reactive to a proactive paradigm, the system delivers tangible benefits:

• Enhances Public Health (SDG 3): Prevents the accumulation of hazardous waste.
• Builds Sustainable Cities (SDG 11): Improves the efficiency and effectiveness of essential municipal services.
• Promotes Responsible Production (SDG 12): Reduces waste and optimizes resource allocation.
• Fosters Innovation and Resilient Infrastructure (SDG 9): Demonstrates a scalable, low-cost, high-impact technological solution.
• Contributes to Climate Action (SDG 13): Reduces emissions from collection vehicles through optimized routing.

Future work will focus on extending the ProWaste pilot to other cities and incorporating additional waste streams, such as recycling and compost. By integrating with vehicle-routing solvers and life-cycle cost analyses, ProWaste can evolve into a comprehensive decision-support platform, further accelerating progress towards sustainable, intelligent, and healthy urban environments for all.

Analysis of Sustainable Development Goals in the Article

1. Which SDGs are addressed or connected to the issues highlighted in the article?

The article on the ProWaste system for urban waste management addresses several Sustainable Development Goals (SDGs) by tackling issues related to public health, environmental pollution, urban living conditions, and technological innovation. The following SDGs are directly connected to the problems and solutions discussed:

• SDG 3: Good Health and Well-being: The article explicitly states that overflowing waste collection centres “threaten public health” through “odour, and leachate.” It also mentions that improperly managed waste can release “toxic fumes and pollutants into the air, causing respiratory disorders and other medical conditions.” The ProWaste system aims to mitigate these health risks by ensuring timely waste collection.
• SDG 6: Clean Water and Sanitation: The article highlights the environmental risk of waste centres polluting “groundwater and soil,” which directly relates to the goal of ensuring clean water and sanitation. The prevention of leachate through proactive management helps protect water resources.
• SDG 9: Industry, Innovation, and Infrastructure: The core of the article is the introduction of “ProWaste, an end-to-end Internet-of-Things and machine-learning platform.” This represents a significant technological innovation (IoT, ML) applied to build resilient and sustainable infrastructure for waste management. The system’s design, using “low-cost sensors” and a “transferable” architecture, promotes accessible and modern infrastructure.
• SDG 11: Sustainable Cities and Communities: This is the central SDG addressed. The article focuses on “Urban waste-collection centres (WCCs)” in “rapidly growing smart cities.” The entire project is framed as a solution to make cities more sustainable by improving “municipal solid waste management” and reducing the “adverse per capita environmental impact of cities.”
• SDG 12: Responsible Consumption and Production: The article deals with the downstream effects of consumption patterns, specifically waste generation and management. By creating a system for the “environmentally sound management of… all wastes,” it directly contributes to this goal. The mention of extending the system to “recycling or composting streams” further aligns with the goal of reducing waste.
• SDG 13: Climate Action: The article notes that “Proper waste management can reduce… climate change problems.” Unmanaged waste in landfills is a major source of methane, a potent greenhouse gas. By optimizing waste collection and management, the system implicitly contributes to mitigating climate change.

2. What specific targets under those SDGs can be identified based on the article’s content?

Based on the issues and solutions presented, several specific SDG targets can be identified:

1. Target 11.6: By 2030, reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management.
   • Explanation: This is the most directly relevant target. The ProWaste system is designed precisely to improve “municipal… waste management.” It also uses the “Air-Pollution_AQI” as a key predictive feature, demonstrating special attention to air quality.
2. Target 3.9: By 2030, substantially reduce the number of deaths and illnesses from hazardous chemicals and air, water and soil pollution and contamination.
   • Explanation: The article links overflowing waste to “public health” threats, “toxic fumes,” “pollutants,” “leachate,” and “respiratory disorders.” The system’s goal of preventing overspill directly addresses the reduction of illnesses caused by waste-related pollution.
3. Target 12.4: By 2020, achieve the environmentally sound management of chemicals and all wastes throughout their life cycle… and significantly reduce their release to air, water and soil in order to minimize their adverse impacts on human health and the environment.
   • Explanation: The ProWaste platform is a tool for achieving the “environmentally sound management of… all wastes.” It aims to prevent the release of pollutants (“leachate,” “odour”) into the environment, thus minimizing adverse impacts.
4. Target 9.4: By 2030, upgrade infrastructure and retrofit industries to make them sustainable, with increased resource-use efficiency and greater adoption of clean and environmentally sound technologies…
   • Explanation: The article proposes an upgrade from “reactively” scheduled maintenance to a proactive, data-driven system. This is a “clean and environmentally sound technology” (IoT and ML) that increases resource-use efficiency by reducing “unnecessary inspections” and optimizing collection routes.
5. Target 6.3: By 2030, improve water quality by reducing pollution, eliminating dumping and minimizing release of hazardous chemicals and materials…
   • Explanation: The system’s ability to prevent “leachate” from waste centres directly contributes to reducing the pollution of “groundwater and soil,” thereby improving local water quality.
6. Target 12.5: By 2030, substantially reduce waste generation through prevention, reduction, recycling and reuse.
   • Explanation: While the primary focus is management, the article states the architecture “can be extended to recycling or composting streams.” This provides the foundation for improved sorting and recycling, which are key to reducing the final volume of waste sent to landfills.

3. Are there any indicators mentioned or implied in the article that can be used to measure progress towards the identified targets?

Yes, the article mentions and implies several quantitative and qualitative indicators that can measure progress towards the identified targets.

• For Target 11.6 (Waste Management & Air Quality):
  • Proportion of municipal solid waste collected: The article aims to “eliminate missed pickups” and prevent “tons of waste going uncollected.” Progress can be measured by tracking the percentage of waste successfully collected against the total generated. The article provides baseline data for Bangalore, such as “total waste collected” (Fig 1c).
  • Air Quality Index (AQI): The article explicitly uses “Air_Pollution_AQI” as a key input variable and a predictor of WCC criticality. Monitoring the AQI in areas around WCCs can serve as a direct indicator of the environmental impact on air quality.
  • Weekly Waste Accumulation: This is identified as one of the three most important features for predicting criticality. Tracking this metric provides a direct measure of waste management effectiveness at a local level.
• For Target 3.9 & 6.3 (Pollution Reduction):
  • Incidence of Leachate and Odour: The system aims to prevent “overspill, odour, and leachate.” A reduction in reported incidents of these issues would be a clear indicator of reduced soil, water, and air contamination.
• For Target 9.4 (Technological Upgrades & Efficiency):
  • System Accuracy: The model’s performance, achieving “>99% accuracy on a hold-out test set,” is a key indicator of its reliability.
  • Reduction in Inspections: The article claims ProWaste can “reduc[e] on-road inspections,” which is a measurable efficiency gain.
  • Time Since Last Maintenance: This is another of the top three predictive features. Optimizing this duration is a key performance indicator for the system’s operational efficiency.
• For Target 12.5 (Recycling):
  • Wet Waste Ratio: The article presents data on the “wet waste ratio” (Fig 1a). An increase in the separation and processing of wet waste (for composting) would be an indicator of progress towards this target, especially if the system is extended as proposed.

4. Table of SDGs, Targets, and Indicators

SDGs	Targets	Indicators
SDG 3: Good Health and Well-being	Target 3.9: By 2030, substantially reduce the number of deaths and illnesses from hazardous chemicals and air, water and soil pollution and contamination.	Air Quality Index (AQI) used as a predictive feature. Reduction in incidents of “odour, and leachate.” Prevention of “toxic fumes and pollutants” that cause “respiratory disorders.”
SDG 6: Clean Water and Sanitation	Target 6.3: By 2030, improve water quality by reducing pollution, eliminating dumping and minimizing release of hazardous chemicals and materials.	Prevention of “leachate” from overflowing waste centres. Reduced risk of “groundwater and soil” pollution.
SDG 9: Industry, Innovation, and Infrastructure	Target 9.4: By 2030, upgrade infrastructure and retrofit industries to make them sustainable… with greater adoption of clean and environmentally sound technologies.	Deployment of the ProWaste IoT and machine-learning platform. Use of “low-cost sensors” and a “mobile app” for data-driven management. Efficiency gains measured by reduced “on-road inspections.”
SDG 11: Sustainable Cities and Communities	Target 11.6: By 2030, reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management.	Improved rate of municipal waste collection (aiming to “eliminate missed pickups”). Data on “total waste collected” and “weekly waste build-up.” Monitoring of “Air-Pollution_AQI” as a key system feature.
SDG 12: Responsible Consumption and Production	Target 12.4: Achieve the environmentally sound management of… all wastes throughout their life cycle. Target 12.5: By 2030, substantially reduce waste generation through… recycling.	Implementation of a proactive system for “environmentally sound management” of solid waste. Data on “wet waste ratio” indicating separation for composting. System’s potential to be “extended to recycling or composting streams.”
SDG 13: Climate Action	Target 13.2: Integrate climate change measures into national policies, strategies and planning.	Implicit reduction of greenhouse gas emissions (methane) from landfills through improved waste management, which reduces “climate change problems.”

Source: nature.com