This paper presents the implementation of an autonomous transportation vehicle as a kinematic test platform for navigation sensors and algorithms. The repeatability of the vehicle trajectory in an outdoor environment with GNSS-visibility will be shown. It is at the same level as most state of the art outdoor kinematic test sites. In addition, a first experiment in estimating the system’s accuracy in areas without GNSS will be demonstrated using a motion model. This model is integrated into a linearized extended Kalman filter framework with autonomous sensors, i.e. IMU and on-board odometry, as control and observation data, respectively.
|