Regresar

Kalman filter implementation in a working cell to classify parts that are in motion

Abstract:

In this article, the efficiency of Kalman's filter to classify moving parts is shown. A SCARA robot working cell was designed and built to accomplish this task. The main goal is to classify parts that are being carried on a conveyor belt at constant speed according to their shape or color. The operation of the Kalman's filter is described, as well as the mathematical equations to develop the algorithm. Tests to determine the pbkp_rediction error are performed, and results from the implementation of Kalman's filter are compared to the ones obtained with equations from uniform rectilinear motion. Additionally, considering the parameters of the working cell, a time optimization analysis, and a production increase analysis are performed, comparing between a conventional classification system and the implemented system with Kalman's filter.