Stereo-vision-aided inertial navigation for unknown indoor and outdoor environments

Reliable knowledge of position and orientation is a prerequisite for many applications regarding unmanned navigation, mapping, or environmental modeling. GNSS-aided inertial navigation is the preferred solution for outdoor applications with the well-known weaknesses and limitations. In order to overcome these, alternative technologies have to be investigated and developed. Therefore a stereo-vision-aided inertial navigation system is presented which is capable of providing real-time local navigation. It neither needs a priori knowledge of the environment nor external references. A method is described to reconstruct the egomotion of a stereo camera system aided by inertial data. In turn, the extracted egomotion is used to constrain the inertial sensor drift. The optical information is derived from natural landmarks, extracted, and tracked over consecutive stereo image pairs. Using inertial data for feature tracking effectively reduces computational costs and at the same time increases the reliability. Additionally, the measured gravity serves as a vertical reference, stabilizing the navigation solution. An Integrated Positioning System (IPS) was developed and tested on real data. Vision data and inertial measurements are fused within an error state space Kalman filter. To do so, aspects like precise and reliable synchronization, calibration and registration have to be considered. IPS was evaluated for accuracy, robustness, and repeatability in a combined indoor and outdoor environment. The outstanding performance of the system is shown on a pedestrian navigation task.