TRASH2CASH: An AI-Assisted Mobile Application for Incentivized Waste Management and Sustainable Resource Recovery

Abstract

Improper waste segregation and inefficient household waste collection remain major challenges in rapidly urbanizing cities, placing considerable pressure on existing municipal infrastructure, contributing to environmental pollution, and leading to the loss of valuable recyclable resources. In many households, dry, wet, and electronic waste are still mixed together, making recycling less effective and increasing landfill processing costs. Although source-level segregation is essential for sustainable waste management, many citizens lack the motivation to participate due to a lack of direct incentives. To address these challenges, this paper presents TRASH2CASH, a cross-platform mobile application developed using the Flutter framework that encourages responsible waste disposal through financial rewards. The application allows registered users to submit waste collection requests by uploading photographic evidence paired with geotagged location data. An AI-powered image analysis module automatically processes the uploaded images to classify the waste type and identify constituent objects, thereby reducing manual data entry and improving classification accuracy. Based on administrator-defined per-kilogram rates, the system estimates the monetary value of the submitted waste. Following verification by a municipal administrator through a dedicated dashboard, approved payments are credited to the user's in-app wallet in real time. Beyond waste collection, the platform integrates a community-driven reuse marketplace for item donations, a grievance management module for reporting unresolved collection issues, and an environmental impact tracker that visualizes landfill diversion metrics and estimated 〖"CO" 〗_2 emission reductions via interactive charts. To simplify deployment and minimize infrastructure overhead, application data is stored locally using a JSON-based file storage system managed through a centralized state management layer, providing complete create, read, update, and delete (CRUD) functionality without requiring a continuously available backend server. Preliminary evaluation of the prototype indicates that integrating AI-assisted waste classification with a gamified reward system significantly increases citizen participation in source-level waste segregation, promotes recycling, and effectively supports sustainable resource recovery.