Evaluation of Accelerometers as Inertial Navigation System for Mobile Robots

Abstract

This article describes the tests and results of the evaluation of an inertial navigation system for a mobile robot, as a mechanism for supplying additional information to the self-localization process. The mobile robot under study has differential steering, a built-in odometric system based on optical encoders for each wheel, and an on-board processor that calculates the position relative to any start point. The inertial system consists of an X-Y dual axis accelerometer, to calculate the position through double integration. The errors of the models are examined and included into the sensors and mobile robot models to validate the estimated position. Many movement conditions like straight and circular paths, including collisions, are identified with both odometric and inertial systems to obtain a more accurate value of the relative position. Results are presented for these conditions, showing that information given by a low cost navigation system based on encoders and accelerometers, with the proper signal processing, can reduce long-term errors and drifts.