Prisma Satellites - GPS Calibration

(Duration of 4 days)

Abstract
The first of the DLR's primary experimental slots consists of a calibration activity dedicated to the characterization and tuning of the GPS-based navigation system. SSC will also participate in this activity. During this experiment the GPS-based navigation system will be closely monitored under various circumstances which will be encountered during the mission. These will include orbit maneuvers with and without Delta-V attitude guidance, as well as different attitude pointing modes on MAIN and TARGET.

Description
The overall GPS Calibration campaign is split in dedicated slots of ca. 12-24 hours time interval (ca. 4 slots of 12 hours each). The main goal is to consolidate the real-time navigation filter parameters.
Each slot stimulates the GPS-based navigation functionalities in an incremental way, with increasing level of complexity.
First clean GPS operations will be performed with quasi-zenith GPS antenna pointing on MAIN and TARGET. An ideal relative eccentricity/inclination vector separation shall be guaranteed during this phase. This will ensure the passive safety of the formation and will not require the execution of orbit and attitude control maneuvers. The subsequent GPS Precise Orbit Determination (POD) process will validate the GPS HW/SW system and support the characterization and tuning of the real-time navigation system. Both nominal and redundant GPS branches will be used during this phase.
Second attitude maneuvers will be performed to drive the MAIN and TARGET spacecraft to different representative attitude pointing modes. Again post-facto POD results will support the interpretation of the obtained results.
Third manual orbit control maneuvers shall be performed, of different size (ca. 1-2 mm/s, 5 cm/s) and direction (ca. cross-track, radial/along-track), no Delta-V attitude guidance shall be performed at this stage. On-ground POD will support the calibration of the executed maneuvers.
Finally manual or autonomous orbit control manuevers shall be performed combined with the Delta-V attitude guidance functionality. A repetition of the previous maneuvers could be envisioned for cross-validation.
The inter-spacecaft separation during the GPS Calibation campaign will be limited to short range GPS navigation (ca. < 2 km).

Upon completion of the GPS calibration, the PRISMA formation can be considered safe and the commissioning phase is complete.

Reports:

First results of the GPS Calibration

DLR has been amazingly fast with evaluation of the first orbits of data. They have evaluated the results of the GPS based relative navigation and suggested a first tuning of the filters to improve the performance. The estimated performance before this tuning is still in the order of meters, specially during propagation, and Target attitude changes. In their own words:

"the default settings provide an unbalanced filter behavior. Due to the small measurements std and the very large empirical accelerations std, the filter gives too much trust to the measurements and does not follow the dynamics. As a consequence each disturbance, un-modelled effect and bias is absorbed by the empirical accelerations (e.g. attitude rotations) which grow to excessively high values. At each propagation phase the wrongly estimated empirical accelerations act as a force and drive the navigation error wild."
"I have tuned the filter and derived new settings which provide a more balanced behavior of the navigation during both Zenith pointing and large attitude rotation phases with frequent data gaps (!). [...] The tuned navigation error is also smooth and is shown to be at the cm level during Zenith pointing, and at ca. 0.5 m during attitude rotations." (Simone d'Amico)

The plot below (click on it for larger version) shows the estimated errors of the onboard navigation (after retuning and running offline) in black, and the covariances of the Precise Orbit Determination (POD) in red. The POD is the orbit determination that DLR runs continuously on ground to give the best possible estimate of the trajectory. It is our reference for all experiments as it is better then the onboard navigation.

Written by

Ron Noteborn

2010-08-19 / 06:26:52

DVS of Moon and Earth before GPS calibration

Before the GPS calibration was started we were still allowed to change the attitude of Mango. This meant that we could test more pointing attitudes and one of these was to point at the moon. This was done together with the DVS camera which meant that we could see the moon rise over the horizon. Here below you can see the result. Due to the wide angle of the camera the moon looks quite small, but one can clearly see the phase it is in. After reaching the moon, we reorient back to the normal attitude, to be ready for the upcomming GPS calibration experiments. With the camera on we can see it sweeping over Earth, from limb to limb. The sequence timing is ~7s between each frame

Written by

Erik Clacey

2010-08-17 / 07:45:36

GPS Calibration Started!

Since yesterday we have been working on the GPS Navigation. Now that we have the two satellites flying in the vicinity of each other and Target is fully stabilised, DLR gets a chance to evaluate their navigation filters. Yesterday was a quiet day where we flew both satellites with one GPS antenna pointing to the sky, quietly tracking satellites. Today and tomorrow we make it a bit more difficult for the filter: we fire the thrusters to change the relative position and velocity and see that the filters take this in. We just have completed the third manoeuvre in a sequence of 9 thrusts. The first two resulted in the orbit that you can see in the plot below. gpscal_M2

First Mango is quietly orbiting at 200-300m distance, but then we fire the thrusters and make a velocity change of 2 mm/s. That gives a small change. The big change in the orbit, where Mango really flies away, is part of the M2 manoeuvre, which is 2cm/s. It was performed at 17:05 UTC. We're currently flying over 1200 m away from Tango. The Intersatellite Link is holding steady and the two satellites keep up their relative navigation. The last manoeuvre (M9) is planned for 11:18 UTC tomorrow morning. We will then repeat the whole sequence in a slightly different mode (where the satellite will turn itself to point a thruster in the desired direction).

Written by

Ron Noteborn

2010-08-17 / 01:46:53