PROTOTIPE SISTEM PENYARING SAMPAH PINTAR BERBASIS IOT DENGAN KENDALI JARAK JAUH UNTUK OPTIMALISASI KEBERSIHAN SALURAN AIR

Abstract

The increase in waste volume along with urbanization growth poses challenges in waste management. The sorting process, which still relies on public participation, often faces obstacles in distinguishing between organic and inorganic waste, leading to waste accumulation especially in waterways which can obstruct water flow and increase the risk of flooding. This study aims to design and implement a prototype of a smart waste filtration system based on the Internet of Things (IoT) with remote control to optimize waterway cleanliness. The system uses an ESP32 as the controller for two continuous rotation servo motors on the conveyor and an ESP32-CAM as the waste detection device using the Faster Objects, More Objects (FOMO) model developed through Edge Impulse. The detection results control the sorting servo to separate organic and inorganic waste, while sorting data is sent in real-time to ThingSpeak via WiFi. Testing was conducted on six types of waste: cups, bottles, plastic bags, leaves, twigs, and banana peels, with each object tested 15 times. The detection success rates were 66%, 86%, 66%, 33%, 66%, and 33%, respectively, yielding an average detection success rate of 58%. Actuator testing showed that servo.write(85) produced a duty cycle of 6.9% and a speed of 68.1 RPM on the transport conveyor, while servo.write(75) produced a duty cycle of 6.4% and a speed of 74.0 RPM on the detection line conveyor. The sorting servo at servo.write(40) produced a 150° movement angle, suitable for directing organic waste. The results show that the prototype is capable of integrating object detection, automatic sorting, and IoT-based monitoring, although model optimization and additional dataset expansion are still needed to improve detection accuracy.