Desarrollo de un algoritmo de auto ensamblaje y generación de movimientos para un robot modular (tipo cadena) utilizando un sensor Kinect

Abstract

This article presents the development of an algorithm that enables the self-assembly of the EMERGE modular robot in three different morphologies. The Kinect sensor was used as the primary data acquisition source, and the information provided by this sensor underwent digital signal processing to individually detect the modules in a workspace. The individual detection of the modules assigns a location to each of them in the workspace, which served as the basis for developing a trajectory that executes movement tables suitable for the self-assembly process in chain, L, and T morphologies. The algorithm was developed in Python, which included a user interface to visualize the real-time processing of computer vision and detect the necessary trajectory ranges for the robot's self-assembly in the selected morphology. Comparative tables were also created to analyze the robot's movements. It was found that, to minimize significant deviations and rotations with the highest possible precision, it was necessary to start with at least three modules.