Indoor localization by using stereoscopic vision, odometry, and the Kalman Filter

Abstract

In order to obtain the location of two robotic platforms within a controlled environment in terms of color and brightness, an algorithm based on Kalman Filter has been developed. This algorithm considers the platform's odometry as the process and stereoscopic vision as the external observer. The tracking is implemented through stereoscopic vision and color object detection. Finally, the algorithm was tested by using Robotino® platforms from Festo, showing a good performance even with the loss of the observer.