An identified LPV model for mobile robots navigation with audio features

Abstract

Non-speech audio is becoming more attractive to be used as features to mobile robots navigation in industrial environments. In this paper authors present their advances in determining robot's position in indoor spaces using as sound sources industrial machines. A novel model is build to locate the robot under different spaces. An identification process is used to obtain the LPV model and it is validated using a real robot. Some uncertainties due to the robot motion and other factors have been taken into account when determining the robot's position and the obtained results demonstrate the validity of the model.