AI-POWERED SUSTAINABLE WASTE MANAGEMENT IN UZBEKISTAN: A VISION-LANGUAGE APPROACH TO SMART SORTING
DOI:
https://doi.org/10.65164/19n68p46Keywords:
smart waste management, vision-language model, edge AI, Raspberry Pi, recycling, circular economy, Uzbekistan.Abstract
Uzbekistan generated 14.8 million tonnes of household waste in 2024, but only 900,500 tonnes (6.1%) were processed by specialized enterprises [2]. At the same time, the country’s Third Nationally Determined Contribution targets 60–65% recycling of solid household waste by 2028 [3]. This paper proposes EcoBin, a camera-first smart recycling MVP for PET bottles and aluminium cans. The design uses a Raspberry Pi 5 controller [9], Camera Module 3 [10], controlled LED chamber, trigger sensor, servo-based three-way gate, Telegram Bot API [11], and hybrid edge/cloud vision-language inference. Unlike fill-level-only smart bins, EcoBin performs item-level acceptance, rejection, routing, reward messaging, and metadata logging. The paper contributes a buildable hardware/software specification, an explicit cycle workflow, and a costed B2G/B2B/CSR pilot model grounded in Uzbekistan’s waste-policy context [2], [3], [8], [16].
