An Approximate Bayesian Method for Simultaneous Localisation and Mapping

This paper describes a Bayesian formulation of the Simultaneous Localisation and Mapping (SLAM) problem. Previously, the SLAM problem could only be solved in real time through the use of the Kalman Filter. This generally restricts the application of SLAM methods to domains with straight-forward (analytic) environment and sensor models. In this paper the Sum-of-Gaussian (SOG) method is used to approximate more general (arbitrary) probability distributions. This representation permits the generalizations made possible by Monte-Carlo methods, while inheriting the real-time computational advantages of the Kalman filter. The method is demonstrated by its application to sub-sea field data. The sub-sea data consists of both sonar and visual information of near-field landmarks. This is a particularly challenging problem incorporating diverse sensing modalities, amorphous environment features, and poorly known vehicle dynamics; none of which can be easily handled by Kalman filter-based SLAM algorithms.