Automated Facade Cleaning Robot using Reinforcement Learning model

Abstract

The maintenance of building facades has traditionally relied on manual labor, which is both timeconsuming and potentially hazardous. The development of automated systems for such tasks offers significant improvements in safety and efficiency. This paper introduces an innovative Automated Facade Cleaning Robot that utilizes reinforcement learning to navigate and clean vertical surfaces independently. Equipped with an Arduino UNO microcontroller, ultrasonic sensors for precise navigation, and a L298N motor driver for motor control, the robot integrates its learning techniques with practical data collected by the environment. The system’s cleaning mechanism, powered by DC motors ensures effective debris removal using rotating brushes enhanced by a water distribution system. The robot is not only capable of working on ground conditions but can also work efficiently on vertical surfaces with the help of its brushless motors. Initial testing shows that the robot can significantly reduce the time and labor required for facade cleaning, promising a new direction in building maintenance technology. Future work will focus on optimizing the learning algorithm and improving the water usage efficiency for better results.