============================================================================
===

Questionnaire concerning GLONASS biases:
----------------------------------------

Your institution:
Contact email address(es):

The GLONASS clock corrections you generate are consistent to:
[x] pseudorange
[x] carrier phase
Please mark both fields if indicated.

Clock corrections are generated at intervals of:
300 s in the main solutions + 30 s in the additional solutions


Handling of GLONASS interfrequency code biases (for ionosphere-free LC):
[x] one bias parameter for each station and satellite combination/pair [ ]
one bias parameter for each station and frequency channel combination/pair [
] other configuration:
Time resolution:
One day. But the interval covered is about 2 weeks. 


Description of datum definition (or reference) for GLONASS code biases:
P1+P2 observations are used according to IGS requirements in the
determination of reference. The procedure is complicated enough and it is
done in several steps. 
Initially it is based on the real ambiguities resulted from daily PPP
solutions for every session (nominally pass) for about 180 stations. In the
preparation of PPP solutions all phase data are adjusted to code
measurements so the main assumption is that inconsistency (or ambiguity) is
caused mainly by receiver-satellite bias.  All the adjusted ambiguities
covering 5-days span (GLONASS+GPS) are processed together using LSM forming
all possible double differences (DD) of the ambiguities. Data weighting is
back proportional of DD-session duration. Number of the parameters to be
solved usually exceeds 9000. All station GPS\GLONASS offsets are adjusted
together with biases. For each pair Stations-Satellite only one bias is
estimated preferably P1+P2. The PPP ambiguities derived from other
combinations are rejected. Weighted sum of the biases only is fixed to 0 for
each SC. The same is done for each station. As a result the file generated
containing one string per each satellite and station: actual time period,
number of DD, bias, bias noise (noise of DD which clearly is affected by
unstable stations) and measured combination (P1+P2 or C1+P2 or ...). In
order to reduce the influence of unstable stations some of them (~10) are
processed separately.
On the second step all biases are recalculated following the requirement to
make clock parameters consistent with P1+P2 data. All the biases are
recalculated wrt to mean P1+P2 receiver. For this purpose all the satellites
are processed independently. The correction to all satellite-related biases
is counted from the condition that weighted sum of P1+P2 only bases should
be 0. 
Biases derived from other combinations are corrected only. Then similar
procedure is performed for the stations. The process is iterative. The
corrected biases are stored in the file. 
On the third step these files covering 14 time span are averaged taking into
account fluctuations of mean P1+P2 receiver!!!
On the last step these biases are merged with C1-P1 and C2-P2 biases in
order to generate the biases for all possible iono-free combinations. C1-P1
and C1-P2 biases are determined in the routine daily processing and then
they are averaged on two weeks interval just before the combinations with
station-satellite biases.

How many conditions/constraints are imposed for the bias datum definition?
See above.


How do you select GLONASS code observable types (to be used for a specific
station) if P1 and C1, or P2 and C2 are available?
P1 and P2 are always preferably. C1 and C2 may be used in case of lack of P1
or P2 data. Anyway in one session only one type of data may be used P1 or
C1, P2 or C2. Generally any combination may be used since we have consistent
biases for all of them.


Do you generate GLONASS clock corrections on a regular/operational basis?
[X] Yes. For both 300 and 30 s data.
[ } No, but test results (in clock-RINEX) would be available for:


Recommendations for PPP users of your GLONASS clock product:
To use our bias files. 


Do you have corresponding GLONASS bias results available (in ASCII form)?
[X] Yes. In order to assess GNSS performance we firstly monitor the
difference between our final and broadcast data. Processing the difference
for the last 24-days period we find the biases between final and broadcast
data. So we have as least two sets of biases: one set is consistent with our
final data and another one with the broadcast data. 
[ ] No

Are there any specialties concerning your GLONASS clock product?
We try to keep final GLONASS clocks consistent with the broadcast clocks. In
other words we try to follow GLONASS system time. There is no visible
satellite clock jumps between consequent daily solutions. 


Extra question: Do you have experience with GLONASS ambiguity resolution?
[X] No
[ ] Yes, please specify:

============================================================================
===