This paper studies the fusion of control inputs and IMU data for developing a kinematic bicycle motion model for the KTH smart mobility lab small-vehicles-for-autonomy (SVEA) platform. This motion model is filtered with relative pose estimates between a camera and fiducial markers, using both an extended Kalman filter and a particle filter. The developed motion models and filters are implemented on SVEA vehicles and are tested in the smart mobility lab. Pose estimates from the motion model and filters are compared against ground truth, determined by a motion capture system with sub-millimeter accuracy. The results presented provide the necessary base for development of automated vehicle control technologies on the SVEA platform with perception based on the detection of fiducial markers.Coble, KyleMahajan, AkanshuKaul, SharangSingh, H. P.
Motion Model and Filtering Techniques for Scaled Vehicle Localization with Fiducial Marker Detection
Advs in Intelligent Syst., Computing
2021-07-31
15 pages
Article/Chapter (Book)
Electronic Resource
English
ARTag, a fiducial marker system using digital techniques
IEEE | 2005
|Vehicle component with image sensor aimed at fiducial marker
European Patent Office | 2025
|VEHICLE COMPONENT WITH IMAGE SENSOR AIMED AT FIDUCIAL MARKER
European Patent Office | 2023
|Vehicle component with image sensor aimed at fiducial marker
European Patent Office | 2023
|