Sensor fusion based on multi-self-organizing maps for SLAM

This paper proposes the use of topological maps in order to provide a method of SLAM feature, based on sensor fusion, that treats better the problem of inaccuracy of the current systems. The contribution of the work is in the algorithm that uses multiple sensory sources, multiple topological maps, to improve the estimation of localization, in order to be as generic as possible, so the same is valid for both internal and external environments (structured or not). When this is made with sensors of clashing characteristics we can obtain better results, because something not perceived by a sensor might be perceived by others, so we can also reduce the effects of error measurement and obtaining a method that works with the uncertainties of the sensors. A simulator was developed to validate the proposed system, through a series of tests with a set of real data. The results show the robustness of the system in relation to the sensorial imprecision and to the gain in predicting the robot's location, resulting in a more appropriate treatment to the errors associated with each sensor.