Representing partial and uncertain sensorial information using the theory of symmetries

In this paper, we propose a general model for representing sensorial information and its uncertainty, the Symmetries and Perturbation model (SPmodel). In it, the intrinsic partiality of geometric information is represented in terms of the symmetries of the involved geometric elements. Location uncertainty due to sensor imprecision is represented by means of a local perturbation, expressed in a reference frame attached to the geometric element, with an associated probabilistic model. Using the SPmodel, we develop a method for integrating geometric information that allows to estimate the location of a feature or an object from a set of partial and uncertain observations. The integration mechanism is based on extended Kalman filter theory.