Two-Dimensional Stochastic Projections for Tight Integration of Optical and Inertial Sensors for Navigation

Aircraft navigation information (position, velocity, and attitude) can be determined using optical measurements from imaging sensors combined with an inertial navigation system. This can be accomplished by tracking the locations of optical features in multiple images and using the resulting geometry to estimate and remove inertial errors. A critical factor governing the performance of optical inertial navigation systems is the robustness of the feature tracking algorithm. Robust feature tracking research has focused on developing multi-dimensional feature transformations which are invariant to camera pose variations. In addition, significant effort has been placed into algorithms designed to pair features between images from large sets (e.g., RANSAC). This traditional approach requires large computational resources, especially when presented with imaging situations with sparse, partially obscured, or repetitive features. In this paper, the method of multi-dimensional stochastic constraints is applied to the optical-inertial navigation problem in two dimensional feature space. The resulting navigation system uses inertial measurements to aid the feature tracking algorithm, which results in improvements in robustness and processing speed. The performance of the optical-inertial navigation system is demonstrated using experimental data. The original document contains color images.