During end of May and beginning of June, we conducted the first closed loop experiments with PROX/VBS. The objective was to stabilise Mango on a point on the positive V-bar, 10m from Tango. From here, the docking target can be observed. These experiments were highly successful, although data processing is still ongoing.

The first experiment simply evaluated the functionality of PROX with VBS in closed loop. This meant the commissioning of a new attitude controller, not used yet in the mission as it is dedicated to this experiment. It also meant the use of the tunings developed for the VBS navigation filter, to which the previous open loop tests were dedicated. Mango responded very well to this, and we immediately observed a positioning quality of better then 5cm in cross track and 10cm in along track. Experiments were not always concluded successfully due to absence of VBS sensor data, leading to safe mode.

Further developments in the experiment were dedicated to new sensor software, and a successive application of measures that were expected to increase the positioning performance, including control box settings, control frequency and application of differential thrust. On June 9, the last experiment (for now) was performed, were the positioning error had been reduced to better then 5cm along track, and 3cm cross track. The standard deviations of the positioning error were 1.4, 0.9 and 1.3cm (along, cross and radial).

The advantages of using VBS over GPS in this kind of forced motion experiments became very apparant, as optical navigation is a much more direct measurement of relative spacecraft state. The GPS navigation relies heavily on modelled dynamics to reduce the rather large raw measurement errors to the 5cm level we are used to with GPS Navigation on Prisma. During forced motion (were thrusts are applied once per minute to once per three minutes) the GPS navigation has a really hard time, but still manages to give decimeter accuracy. The optical navigation improves this, as the navigation filter relies less on the dynamics, and the measurement in itself is of higher quality then GPS raw data. The result is less slowly varying biases, making it possible to position the spacecraft center of mass within a large match box.

Below is a video made from the data from the VBS (candidates for LED positions are blue crosses, with the Tango pose drawn in blue when it is available from the VBS. Grey is the reconstruction from oboard GPS and Tango magnetometer/sun sensor based estimation). Initially the spacecraft is controlled by GPS, but then transitions into PROX/VBS. The reorientation to account for camera misalignment can be clearly seen, as well as that the VBS looses track due to the high rotation rate at that moment. The spacecraft is then stationkeeping at 10m distance.

The PROX/VBS experiments took place in the spacecraft local orbit frame (SLO), but in the mean time we have done the first experiments in the Tango body reference frame (SCB). These experiments fix Mango on a point relative to the Tango spacecraft structure, meaning that if Tango moves, Mango moves along. The first experiment placed Mango on 10cm on the -Y panel. This is basically the same point as the SLO experiment, but since Tango has a misalignment with SLO of some 10 degrees, the off positioning from the Vbar point was as large as 2m. Mango indeed succeeded to arrive at this point, at which the FARM mode was activated. FARM has the same functionality as PROX (it can control in Tango SCB frame) but it has extra guidance logic to have a position dependant control box, and can autonomously move towards a predefined docking location. The first experiment again commissioned the software: a new attitude guidance function was to be checked out, which points the VBS camera to the docking target) and the capability of FARM to keep the spacecraft in the Tango SCB was evaluated. The experiment was a success: Mango was over an orbit in the Tango SCB frame, moving from 10m to 8.5m in 1500s and back again. The attitude guidance worked well. We are now waiting for the next opportunity to test this further, and to decrease the distance.

More to follow!