151 
3.2 Metrology experiment 
Simplified Scheme of the Testbed 
A laboratory experiment has been carried out in 
collaboration with the Metrology Institute Colonnetti 
(Torino) to test the performance of the GAIA metrology 
system, with the goal of stabilizing the spacing of the 
optical components to within 100 pm over distances 
of a few meters. To date, laser interferometry is the 
best measurement technology for achieving this 
performance. 
The experiment layout is shown in Fig. 9. It consists of 
a stabilized Nd:Yag laser and two plates, which 
simulate the generic optical elements of GAIA, faced at 
0.5 m (Gai et al, 1997). Each plate is moved by means 
of piezotraslators, while the Nd:Yag laser is stabilized 
with reference to a Fabri-Perot cavity, and the plate 
spacing is stabilized with reference to the wavelength of 
this laser. The resolution required is achieveble only 
maintaining the enviromental disturbances at a 
minimum. For this reason the testbed plates are placed 
in a vacuum bell, placed on a table supported by three 
pneumatic isolators. Note that thermal and acoustic 
effects would change the local refraction index, 
resulting in uncontrolled variation of the optical paths. 
The distances of the reference points between the 
plates are measured by means of Fabry-Perot 
interferometers, and actively controlled to within 100 pm 
(picometers). In a Fabry-Perot interferometer the 
monochromatic light is channeled through a pair of 
parallel half-silvered glass plates, producing circular 
interference fringes. One of the glass plates is 
adjustable, enabling the separation of the plate to be 
varied. The wavelength of the light can be determined 
by observing the fringes while adjusting the separation. 
The basic Nd:Yag laser is a ligthwave mod. 146 
operating at X=1.064 pm, with narrow emission and a 
wavelength stability of ~ 1 part on 10 8 , for a three 
hours observation time. The laser is stabilized with 
reference to an optical cavity, with a ultra-low thermal 
expansion coefficient, by means of the Pound-Drever 
method (reference material: ULE®, CTE = 10' 8 IK). 
The required reference cavity temperature stability is 5T 
< 8-1 O' 4 K 1a over 0.75 s * 3 h time scales. 
The absolute distance measurement principle is the 
following: the absolute distance (s) between two 
markers is obtained from the frequency variation 
measurements Av, in mathematical form we have: 
where c is the light velocity. 
The requirements for GAIA correspond to distance 
stabilization constraints to better then 2 pm. 
The experiment lead to very interesting results: digital 
control kept the average distance variations between 
the mirror pair at the level of 2pm (1a). Moreover, digital 
control guarantees to maintain these performances 
unaltered during longer time scales like in the GAIA 
system (3 hours). 
The sub-nanometric optics stabilization is then 
demonstrated. 
DATA PROCESSING 
In analogy with its predecessor HIPPARCOS, GAIA 
makes use of very highly accurate measurements along 
the scanning direction; this capability, in conjunction 
with the peculiar scanning law according to which the 
same object is observed many times at different 
scanning orientation, allows to reconstruct the stellar 
astrometric parameters with a precision of a few micro- 
arcseconds. In principle, the problem is very much 
similar to that of HIPPARCOS, with the difference that 
the number of stars involved is ~ 10 4 times larger and 
the target error - 10 3 smaller. 
Therefore, the data reduction task has to take into 
account all possible effects, both of physical nature as 
well as numerical, which can introduce an 
indétermination or a systematic error at the level of the 
sought for accuracy (Bucciarelli et al., 1997). We do not 
discuss here this kind of details; instead, we present an 
outline of the principles of the reduction strategy which 
will need to be further developed.