Desarrollo de un sistema inteligente para la gestión de la recolección de desechos sólidos urbanos basado en comunicación inalámbrica

Abstract

The objective of this degree work was to develop an intelligent system for the management of urban solid waste collection based on wireless communication. To achieve this, some requirements have been defined, such as the generation of a wireless network model that considers the dumpsters as nodes connected to a remote graphical interface, the coding of an artificial vision algorithm that identifies the fill level of the dumpsters, and the determining of their location. As a result, an intelligent system coded in Python and processed on a Raspberry Pi 4 Model B was obtained, which by means of an artificial vision model based on optical character recognition (OCR) detects the filling level of the garbage containers and sends this information to a remote graphic interface created on the Blynk IoT platform. The remote graphical interface monitors the status of dumpsters, displays their location and the closest filled bin, as well as allows the user controls access to them. The prototypes are coupled with an MG996R servo motor that locks the door when full and an FC-51 infrared sensor that triggers a voice alert when access is attempted. Functional tests showed that the image processing algorithm has an efficiency of 86.36% when used with dark-colored waste, and the "closest bin" function only proved useful when used in residential areas. Finally, the response time of the door locking system is 28.19 seconds, while that of the alarm system is 20.11 seconds. For future work, it is recommended to maintain constant lighting and test the prototype with a stable internet network.