Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

A sensor-centric EKF for inertial-aided visual odometry

: Kleinert, Markus; Stilla, Uwe


Canalda, P. ; Institute of Electrical and Electronics Engineers -IEEE-:
IPIN 2013, 4th International Conference on Indoor Positioning and Indoor Navigation : Montbeliard, France, 28 - 31 October 2013
Piscataway, NJ: IEEE, 2013
ISBN: 978-1-4799-4043-1 (online)
ISBN: 978-1-4799-4042-4 (print)
8 S.
International Conference on Indoor Positioning and Indoor Navigation (IPIN) <4, 2013, Montbéliard>
Fraunhofer IOSB ()

When appropriate infrastructure is not available, localization of pedestrians becomes a difficult task. This is especially the case in urban or indoor scenarios, where satellite navigation is hindered due to occlusions or multipath effects. A promising alternative is to combine a small, low-cost inertial measurement unit (IMU) with a camera in order to exploit the complementary error characteristics of these devices by simultaneously estimating the positions of observed landmarks and the trajectory of the sensor system with a stochastic filter. This work presents a purely sensor-centric formulation of the monocular simultaneous localization and mapping (SLAM) problem where the reference system is fixed to the platform's coordinate system. As a consequence, the prediction and update step of the employed extended Kalman filter (EKF) are adapted such that the 3D-position of observed landmarks relative to the sensor is propagated in time and the jacobians calculated in the update step depend only on the estimated landmarks. Furthermore, it is demonstrated how the state prediction equations can be modified in order to reduce the computational complexity of the prediction step. Finally, results for simulated datasets as well as for a real indoor dataset are shown in order to assess the performance of the algorithm.