Using onboard data fusion of IMU and GNSS for improvement of scientific rocket flights

Quite often scientific sounding rocket missions, such as the microgravity-mission MAIUS in 2017, require an as accurate as possible attitude determination. Up to now, GNSS and IMU data have mostly been used separately on sounding rocket flights. IMU data have been used for attitude determination solely, while GNSS data provided position and velocity vector of the rocket with a high accuracy but low sample rate. Normally, this data have so far been combined only during post processing. In a recent sounding rocket flight program, the atmospheric physics mission PMWE, MORABA, the Mobile Rocket Base of DLR - German Aerospace Center, was responsible for trajectory and attitude determination as well as for the live data handling on board and the communication with the ground stations. During this mission, a newly implemented algorithm, running on the onboard computer, performed the combination of the highly sampled, but drifting IMU data and of the very accurate, but low sampled GNSS data. Although the GNSS only delivers position and velocity vector, the attitude data could be improved with the help of this fusion algorithm. The used algorithm is based on Kalman filtering and was used for the first time on a MORABA sounding rocket flight. Furthermore, during post-flight processing a data fusion with the measurements of a second GPS-receiver was made. The results show that this technique can also be useful for missions, which require advanced guidance and control. For example, the inclination and orbital accuracy of satellite launchers could be improved as well. In this paper we present on the basis of the flight data of the PMWE sounding rocket mission the results of the on-board as well as of the post-processing fusion, and the improvements which have been gained with this new technique.