Guaranteed Estimation when Integrating Sensor Information into Robot System

Abstract

A robotic system is considered, in particular, a spacecraft for which the issue of combining information from sensors of the navigation systems arises. An inertial navigation system and a star sensor are taken as the main navigation systems. The algorithm of guaranteed evaluation when combining the information of the inertial navigation system and the star sensor is proposed. When solving the problem of evaluation with the help of the minimal filter, the construction of the information set, which is guaranteed to contain the true state of the system, is a computationally difficult task for large dimensional systems. Therefore, instead of the exact construction of information sets, their approximate value is used, that is approximation in the form of convex polyhedron, received by approximation from above of "exact" but set implicitly by the system of linear equations and inequalities of the information set. The result of guaranteed evaluation is compared with the result of evaluation by the Kalman filter.