Development of a Solid Waste Collector Robot for Cleaning in Public Areas

Solid waste management reduces pollution, protects public health, conserves ecosystems, and promotes recycling and circular economies in a sustainable manner. This study analyzes the pollution issues caused by inefficient solid waste management in urban areas, highlighting its negative impact on the environment. In this vein, the T5R-bot collection robot was developed, designed through 3D printing and implemented with an artificial vision model for object detection based on the single shot multibox detector (SSD) and the MobileNetV2 neural network architecture. This system allows for the autonomous identification and collection of up to 12 types of debris, achieving an accuracy of 98% and a mAP of 97.81%. The methodology included the mechanical design of the robot with a rocker-bogie mechanism, ultrasonic sensors for navigation, and a robotic arm with four degrees of freedom. The model was trained with a dataset collected from 2890 images, demonstrating high efficiency in detecting and collecting waste in contaminated environments in public areas. The results confirm the viability of the robot as a tool for improving solid waste management. In addition, the future integration of segregation and adaptive learning capabilities is proposed.