In this paper, practical realization of a previously proposed VDM based integrated navigation for UAVs is presented. This method utilizes the vehicle dynamic model (VDM) fed by autopilot control commands as main process model within the navigation filter and treats IMU, GNSS or other sensory data as observations. Extensive Monte Carlo simulation results utilizing this approach in previous publications suggested significant improvement in navigation accuracy during long GNSS outages, attitude determination for photogrammetry applications, and robustness in different wind profiles. Preliminary experimental results are presented in this paper on positioning autonomy of an in-house fixed-wing UAV with MEMS IMU onboard during 3 minutes of GNSS outage, revealing more than 20 times improvement in positioning autonomy compared to conventional INS/GNSS integration.