An Entropy-Like Approach to Vision-Aided Inertial Navigation

The paper presents a loosely coupled approach for the improvement of the state estimation in autonomous inertial navigation tasks, augmented via image-based relative motion estimation. The proposed approach uses a novel Pose Estimation technique based on the minimization of a Entropy-Like cost function which is robust by nature to the presence of noise and outliers in the visual features. A Indirect Kalman Navigation Filter is used, in the framework of stochastic cloning. The robust relative pose estimation given by our novel technique is used to feed a relative position fix to the navigation filter. Simulations results are presented and compared with the results obtained via the classical Iterative Closest Point approach.