Historic ILRS Normal Point Format

HISTORY - April 1990

It was recommended at the Laser Ranging Instrumentation Workshop held in Herstmonceux in 1984 that normal points should be formed by laser stations very shortly after the satellite pass and transmitted as a form of quick-look data. The formal recommendation at Herstmonceux provided the detailed algorithm for the computation of normal points which has now become widely accepted. The Herstmonceux discussions concluded that it was feasible to produce on-site normal points which could be transmitted as quick-look data, however, existing formats could not accommodate the additional information deemed necessary for normal points.

In 1988, Bendix Field Engineering Corporation (BFEC, then AlliedSignal Technical Services Corporation, ATSC, now Honeywell Technology Solutions Inc., HTSI) made an initial proposal for a quick-look format that would meet the needs for engineering assessment with sampled quick-look data and the scientific needs associated with the use of normal point data [SLR Subcommission of CSTG Newsletter, July 1988]. The proposed format promoted discussion that led to further versions [SLR Subcommission of CSTG Newsletter, June 1989].

The proposed formats were reviewed and discussed at the Laser Ranging Instrumentation Workshop held at Matera in October, 1989. The consensus of the attendees was that the transmission of normal points as quick-look could make significant contributions to their scientific application and a format should be adopted soon. Furthermore, while the format to support engineering assessment was important, many stations were capable of providing their own assessment and would not transmit both sampled and normal point data. As a consequence of these discussions, a resolution was passed that required the adoption of a format that was as close as possible to the MERIT-II format (now known as ILRS fullrate), but satisfied telex line constraints, and enabled more detailed identification of important quantities, such as time tags. The urgency of the adoption was reiterated with the recommendation that an acceptable format be prepared within 45 days by a committee of representatives from Royal Greenwich Observatory (RGO), University of Texas (UT) and Goddard Space Flight Center (GSFC).

A follow up meeting was held at GSFC during the SLR Subcommission of the CSTG meeting in late October 1989. Additional considerations at this meeting were that, at the time, there was no intention to cease the collection and archiving of full-rate data, and so there was no need for significant analyst components in the quick-look data format that were not already in the MERIT-II format. The widespread acceptance of the MERIT-II format was further emphasized by the lack of substantial criticisms of the MERIT-II format in its current use for both full-rate and normal point data. Hence, a committee consisting of A. Sinclair (RGO), R. Kolenkiewicz (GSFC), and B. Schutz (UT) met to discuss details of the format. D. Edge (BFEC) participated in these discussions which led to the adopted format. Comments on the format were solicited and incorporated. H. Linder (GSFC) made significant contributions to the development of the final format.

The adopted format has removed redundant data from the data record, resulting in a header record containing information for the entire record and a data record containing the time-dependent information. Some items that do not exist in the MERIT-II format have been added, but these are not vital to the analyst, and so the data can be readily reformatted to MERIT-II, if required.

HISTORY- March 1997 Revision

The CSTG on-site normal point format was originally intended for the transmission of quick-look data, but now that full-rate data has been phased out (from April 1, 1996), the site-produced normal points which use this format have become the principal SLR data product. Prior to this transition there had been discussion of the need to revise the format, in order to include additional information that would be lost when full-rate data was no longer available. A proposed revision of the format was put forward by V. Husson (ATSC) at the Laser Ranging Workshop in Canberra in November 1994, and was further discussed at the Eurolas meeting in Munich in March 1995. However at the SLR CSTG Subcommission meeting in Bern in December 1995 it was decided not to make major changes to the format at this time, but just to make some minor revisions to a few of the data items. At the SLR CSTG Subcommission meeting in Shanghai in November 1996 a few further minor revisions were agreed. A major revision of the format will ultimately be needed as more stations move towards operation at single photon return level (e.g., SLR2000, KEYSTONE), and more experience is gained with corrections for satellite signature effects, which are mainly seen at single photon levels.

The following changes were agreed upon in Bern in December 1995:

  • Use of the term RMS (root mean square) throughout instead of a mixture of RMS and standard deviation
  • A modification to the wavelength field to permit wavelengths greater than 1 micron
  • A data sequence flag in the data records, to indicate if the data are the original or a replacement set

The following changes were agreed upon at a follow up meeting in Shanghai in November 1996:

  • Merging the calibration method and calibration shift-type indicators in the header record into a single indicator, in order to free a field
  • Clarification and implementation of the existing configuration flag, as some systems have alternative operational configurations according to the satellite being tracked (e.g., use of a PMT or APD detector)
  • Inclusion of a system change indicator, which will be incremented each time there is a significant change to items of equipment or software
  • Decision to put the data sequence flag (agreed in Bern) in the header record instead of the data record (however on consideration it was realized that it is in fact better in the data record, as agreed in Bern; the header record does not contain the time of day of a pass, and so does not uniquely define a pass, and thus it is not possible by comparing header records and data sequence flags alone to remove passes that have been replaced).

Information describing the various configurations of a system and recording changes made to the system will be maintained in log files in a standard format, and these will be held at the data centers and also at the station. The details are described in a separate note, "SLR System Configurations and System Changes" (SLRMail #0048).

The changes agreed at Bern were described in the CSTG SLR Newsletter dated May 1996. In this note we have slightly revised the description given in that Newsletter for the wavelength field, which specifies that a wavelength value of 0001 to 2999 is to be interpreted as being in units of 1.0 nm, whereas we restrict this range of values to 1000 to 2999. The original form is not inaccurate, but does permit, for example, a wavelength of 0.532 microns to be specified as 0532 or 5320.

At the Laser Ranging Subcommission Meeting in Shanghai it was also recommended that Lunar Laser Ranging data should be set in the same format as SLR normal points, in order to make the data more easily available to a wider analysis community, and to encourage more SLR stations to attempt ranging to the Moon. Only a few fairly minor changes are needed to the SLR format in order to accommodate LLR data, and these have been included in this revision.

LLR normal point data does not now explicitly record values of calibration, calibration standard deviation, and calibration shift, so these fields may be zero filled. Individual stations may choose to include this information if available.

A full description of the changes made to this format as of March 1997 was distributed via SLRMail #0047.

HISTORY- April 1999

During the April 1999 ILRS Meeting, the ILRS Normal Point Format became the new official name of this format formerly known as the CSTG On-Site Normal Point Format.

HISTORY- September 1999

The Jaguar Team, a study group of the Data Formats and Procedures Working Group, concluded "that ILRS makes NO RESTRICTION on the minimum number of returns used to generate Normal Points."

FORMAT INTRODUCTION

The ILRS Normal Point and Engineering Quick-Look Format is designed to accommodate both processed, calibration corrected Herstmonceux defined normal point data and sampled raw engineering quick-look data. The format contains fixed length 55-byte records made up of ASCII characters. All record types have the same record length for easier implementation and use.

A quick-look pass message consists of normal point data records and/or raw engineering data records. Normal point data records are preceded by the indicator '99999' and a header record. Raw engineering data records are preceded by the indicator '88888' and a header record. (Five 8's and five 9's were chosen since seconds of day can not be equal to 88888 or 99999.) The same header format is used for both the normal point data records and the raw engineering data records. Therefore, in most cases, if both normal point data records and raw engineering data records are present for a given pass, the header record for each type of data record would be identical. The only exception would be if a pass crosses midnight, and the first normal point data record is not on the same day (and possibly year) as the first sampled engineering data record.

The descriptions of the header record, normal point data record, and sampled engineering data record are contained in the ILRS Normal Point and Engineering Quick-Look Format Definition document.

A quick-look pass message with m normal point data records and n sampled engineering data records followed by a quick-look pass message with x normal point data records is shown below. (Note that the first pass has both normal point and engineering data records, while the second pass has only normal point data records.)

99999
[Header Record]
[Normal Point Data Record 1]
[Normal Point Data Record 2]
...
...
...
[Normal Point Data Record m]
88888
[Header Record]
[Sampled Engineering Data Record 1]
[Sampled Engineering Data Record 2]
...
...
...
[Sampled Engineering Data Record n]
99999
[Header Record]
[Normal Point Data Record 1]
[Normal Point Data Record 2]
...
...
...
[Normal Point Data Record x]

NOTES: (April 1990)

  1. The range and all corrections fields are in two-way light time units of picoseconds for accuracy and consistency.
  2. The time-of-flight data in the normal point data record should not be corrected for atmospheric delay nor to the center-of-mass of the satellite. Hence no fields are included in the format for the quantitities applied, since the corrections have not been applied. The time of laser firing should be give, not the time at the satellite.
  3. It is recommended that fields for which data values are given should be filled with leading zeros, and fields for which no data value is given should be zero-filled also.
  4. If a data value exceeds the allocated field width then enter all 9s for that quantity.
  5. If the time-of-day of a pass crosses 24 hours UTC then the seconds should be given modulo 86400, but the day-of-year remains unchanged at the value of the first normal point and first engineering record of the pass, respectively.
  6. For two-color ranging the two sets of data should be given with separate headers (in which the wavelength is specified)
  7. The calibration system delay data will be given in columns 24-32 of the header record and in columns 37-44 of the engineering record. For systems using external calibration the value in the header will be the value actually applied to the data. For systems using internal calibration it will be a value indicative of the calibration values applied (e.g., the mean of the calibration values). The calibration system delay will be applied by the station to the time-of-flight in the data record, but not applied to the time-of-flight in the engineering record.
  8. If, having changed to this new format, stations should have difficulty in forming normal points for a particular pass or a particular satellite, then the normal point data record can still be used to transmit sampled raw data points instead, but about 50 fairly equally spaced points should be transmitted (although the engineering data record could in principle be used for this purpose, analysts will not be set up to use it).

NOTES: (March 1997)

  1. The original wording of the format (April 1990) uses the expression 'trend function', and this is continued unchanged in this document. This is a short-hand term for a concept originally described in the recommendations for formation of normal points, proposed at the laser ranging workshop in Herstmonceux in 1984. We repeat here the relevant wording: 'Solve for a set of parameters (orbital parameters preferable) to remove the systematic trends of the prediction residuals, not introducing spurious high-frequency signals into the TREND-FUNCTION'.
  2. It is generally agreed that the checksum is now redundant, and in fact whenever it indicates an error it is invariably due to an error in forming the checksum, rather than a data transmission error. Thus stations can if they wish leave this field blank, but if they continue with it then it should still be the sum of integers in columns 1-52, modulo 100. i.e., the new field in column 55 of the header record should not be included.
  3. It is proposed to use this normal point format for LLR data also, which can be accomplished by just a few additional flags, and corresponding changes to the interpretation of some data fields. These changes are described in an appendix to this note.
  4. A flag has been included to permit indication of replacement data releases, should they be necessary. Stations should be aware that replacement data causes problems for data centers and analysts, and should make strong efforts to get the data right first time.

NOTES: (September 1999)

  1. There is NO RESTRICTION on the minimum number of returns used to generate normal points.