Information-Based Georeferencing Of An Unmanned Aerial Vehicle By Dual State Kalman Filter With Implicit Measurement Equations

Georeferencing a kinematic Multi-Sensor-System (MSS) within crowded areas, such as inner-cities, is a challenging task that should be conducted in the most reliable way possible. In such areas, the Global Navigation Satellite System (GNSS) data either contain inevitable errors or are not continuously available. Regardless of the environmental conditions, an Inertial Measurement Unit (IMU) is always subject to drifting, and therefore it cannot be fully trusted over time. Consequently, suitable filtering techniques are required that can compensate for such possible deficits and subsequently improve the georeferencing results. Sometimes it is also possible to improve the filter quality by engaging additional complementary information. This information could be taken from the surrounding environment of the MSS, which usually appears in the form of geometrical constraints. Since it is possible to have a high amount of such information in an environment of interest, their consideration could lead to an inefficient filtering procedure. Hence, suitable methodologies are necessary to be extended to the filtering framework to increase the efficiency while preserving the filter quality. In the current paper, we propose a Dual State Iterated Extended Kalman Filter (DSIEKF) that can efficiently georeference a MSS by taking into account additional geometrical information. The proposed methodology is based on implicit measurement equations and nonlinear geometrical constraints, which are applied to a real case scenario to further evaluate its performance.