ESA's Infrared Space Observatory (ISO) was successfully launched
from the Guiana Space Centre in Kourou on 17 November 1995. Its
requirements in terms of ground-segment preparation were
particularly demanding due to the limited mission lifetime, which
calls for highly efficient operations, the very fast pace of the
ISO observations, some lasting just a few minutes, the severe
pointing requirements which demand sophisticated planning, and
the real-time commanding of the highly sophisticated payload of
four instruments with multiple operating modes from a computer-
generated, automatically executing file.
It was recognised that, given these demanding constraints, a well
thought out approach to overall ground-segment integration,
testing and validation would be required to ensure success. The
approach that was chosen, based on the concept of end-to-end
testing supported by sophisticated instrument simulators, proved
highly effective. As a result, the ISO ground segment was ready
to support all of the mission's operational phases in time for
the spacecraft's launch.
Two different ground-segment configurations have been established to support ISO operations during the four different mission phases:
Control of the mission was to be transferred from ESOC to Villafranca at the end of LEOP, once the ISO operational orbit (24 h geosynchronous) had been achieved, nominally four days after launch. The Villafranca Control Centre has two co-located main elements, the Spacecraft Control Centre (SCC) and the Science Operations Centre (SOC), communicating via a routing and archiving computer known as the Operational Data Server (ODS) and housed in the SCC (Fig.1).
Figure 1. Simplified overview of data flow within the ISO Control
Centre at Villafranca
Aerial view of the ESA complex at Villafranca
The ISO Spacecraft Control Room at Villafranca
The ground segment has been designed to support automatic operation of both the spacecraft and its on-board instruments. Under nominal conditions, satellite commanding is performed automatically from a Central Command Schedule (CCS) that is generated in advance for each ISO orbit. The CCS is based on the concept of 'operations windows', whereby time slots are allocated for all spacecraft and science instrument commanding activities during a given orbit. Typical examples are star-tracker calibration, ranging, station handovers, instrument activation/deactivation, and the scientific observations themselves.
The Central Command Schedule is generated in several consecutive steps from observations entered by Guest and Guaranteed Time Observers and validated at facilities provided at ESTEC and at NASA's Infrared Processing and Analysis Centre (IPAC), before being transmitted to the mission database:
The science and housekeeping telemetry from the spacecraft, stamped with the reception time at the ground station, is transmitted to the SCC, where the housekeeping telemetry is extracted and processed for spacecraft-control purposes. In parallel, the raw instrument (science and housekeeping) telemetry is augmented with telecommand history and ancillary data and transferred to the SOC in the form of Telemetry Distribution Formats (TDFs) across the ODS interface. These real-time TDFs are archived by the ODS and processed in the SOC at the instrument stations (one per instrument) for real-time assessment of the on- board instruments' behaviour (Real-Time Assessment: RTA) and preliminary analysis of the quality of the science data (Quick- Look Assessment: QLA).
The telemetry archived in the ODS is accessed and further processed offline by the SOC through the Off-Line Processing (OLP) system. This processing includes pipeline and interactive analysis and generates the final mission products, which are then archived at the SOC and distributed to the scientific community.
To increase flexibility and cope with potential spacecraft or instrument anomalies, manual commanding of the spacecraft and the instruments is possible by means of dedicated procedures available to the spacecraft and instrument controllers.
The ISO ground segment has been designed and implemented by several different groups under the overall coordination of the ISO Project at ESTEC:
It was recognised early in the ISO Programme that the complexity of the overall ground segment (cf. Fig. 2) would make its integration and validation a very challenging task. An ISO Integration and Test Team (ITT) was therefore established in the summer of 1991 to generate and implement an overall integration and validation plan, based principally on the concept of 'end-to-end' testing, that would ensure the ground system's readiness in time for the observatory's launch.
Figure 2. The ISO computers and Local Area Networks (LANs) at
Villafranca
After a slow start due in part to the unavailability of much of the ground-segment software, the team became fully operational in the second half of 1993. The intervening period was used to establish and refine the integration and validation concept, as well as the basic ground rules and responsibilities. As work progressed, the ITT's membership was broadened to include spacecraft- and science-operations representatives, as well as representatives of the Instrument Dedicated Teams (IDTs). This expansion of the ITT followed quite naturally from the progress made in the ground segment's ability to support the mission.
Once the ground system was functioning, the emphasis shifted towards performance and operational issues. The final ITT System Operation Validation tests and operational simulations provided early feedback on both operational aspects and system performance. In addition, they offered an opportunity for the extensive training of operational personnel and ensured a smooth transition between the validation testing and the ISO pre-launch simulation campaign at Villafranca.
The ITT approach to ground-segment integration and validation was based on the concept of comprehensive end-to-end testing with the real spacecraft, to complement the traditional system-validation and mission-readiness tests (SVTs and MRTs). The traditional approach of SVTs (validation of spacecraft control operations in terms of database contents and procedures) and MRTs (validation of the Control Centre inter-faces with the network) leaves much of the ground-segment 'final tuning' to be carried out after launch. Because a tuning phase lasting for several months would have reduced the scientific return from the mission, the end-to- end tests were conceived as a means to permit dynamic and incremental validation of an overall system against the flight hardware. More and more of the SOC functional elements and the corresponding operational interfaces were included as they became available to reduce the number of problems that, with the classical SVT MRT approach would have been detectable only after launch.
The end-to-end concept adopted consisted of establishing a validated Control Centre (SOC and SCC) communications backbone on which the applications software was incrementally integrated and validated as and when available.
This concept called for the availability of a representative simulator that would be capable of reproducing the in-orbit behaviour of the spacecraft and its instruments. The requirement to support full-scale testing of the SOC system in both stand- alone and integrated SOC/SCC modes was beyond the capability of the test tools either available or foreseen at that time, and this led to the development of a new Integrated Instrument Simulator (IIS), starting in the second half of 1993. The IIS, development of which was completed in approximately seven months, has proved to be an essential tool in the successful validation of the overall ground segment.
The acquisition of solid 'pre-launch operational experience' was invaluable for a short-lived observatory-type mission such as ISO (planned mission lifetime 18 months), as the highly successful in-orbit-performance verification campaign of the last months has confirmed.
The Villafranca ground segment's validation was based on two levels of testing:
This first level of testing would demonstrate the ground segment's ability to operate the spacecraft properly via the automatic schedule (CCS), including flight-dynamics operations, as well as through manual commanding.
The objectives of the first level of testing were met in the first end-to-end test (EEO) during March/April 1994. The objectives of the second level were largely achieved in the second (EE1) and third (EE2) end-to-end test series conducted from December 1994 through February 1995 and in April/March 1995, respectively. Although EE2 was originally conceived as an 'all bugs fixed' repetition of EE1, actual schedules and programmatic constraints forced EE1 and EE2 to be complementary in the sense that EE1 could only be used to validate about half of the envisaged instrument operational modes, leaving EE2 to validate the rest. The remaining ground-segment functionalities, mostly involving SOC/SCC and SOC-internal, non-real-time interactions, were pre-tested in the period between EE2 and the beginning of the simulation campaign at Villafranca. The final validation was achieved during an intensive set of joint SOC/SCC simulations in July/August 1995.
The ground-segment elements involved in the various EE tests are shown in Figure 3.
Figure 3. The ISO ground-segment elements involved in the various
end-to-end tests and the extended simulations
EEO
The first end-to-end test sequence was performed twice:
Each test lasted five days and was performed using the IIS. Both tests were successful and the objectives set for this first level of testing were achieved. A total of 53 Software Problem Reports (SPRs) were raised, all of which were classified and prioritised according to their importance for the subsequent EE tests.
EE1
This second set of end-to-end tests was carried out on six
consecutive days in February 1995, each day simulating a typical
ISO orbit. The Villafranca Control Centre was connected directly
to the ISO flight model at ESTEC in its checkout environment. The
Test and Simulation Assembly provided simplified in-orbit
conditions for the observatory's Attitude and Orbit Control
System (AOCS), including provision of a single guide star for the
simulation of spacecraft manoeuvre execution. The first three
orbits were dedicated to the validation of performance-
verification observations, the last three to the testing of
observations in standard instrument modes of operation (defined
through Astronomical Observation Templates: AOTs).
ISO Dedicated Computer System (IDCS) display showing the ISO
AOCS top-level mimic and, in the background, alphanumeric windows
for spacecraft control
Deficiencies were highlighted in several areas during the EE1 testing, including: connectivity problems between Villafranca and ESTEC, database inconsistencies, and problems in both operational and support software. A total of 91 SPRs were raised, but many of these problems were already solved prior to starting EE2.
A great deal was learned in EE1 from an operational point of view, particularly regarding the ground segment's ability to recover from errors, the manual operation of the spacecraft and instruments, and the resumption of automatic commanding from the CCS after an interruption.
EE2
The EE2 testing consisted of running seven simulated orbits
from the ISO Control Centre in Villafranca with the ISO flight
model at ESTEC. All orbits were dedicated to testing the various
instrument operational modes (defined through the AOTs), but
included some additional flight-dynamics tests. This entire
series of tests, including rehearsals, dry runs, and the actual
tests with the satellite, was completed in just five weeks,
starting at the end of March 1995.
In the EE2 tests themselves, emphasis was put on the operability of the system, particularly on the SOC side. Several errors were detected and a total of 183 SPRs were issued, and addressed in order of priority.
At the end of EE2 it was concluded that, although functionally ready to support performance verification and routine operations of the ISO mission, from an operational point of view there were still major deficiencies in the system. These were rectified before the start of the extended simulations in July 1995 thanks to the dedication of all groups involved involved in the ground segment's preparation.
Extended simulations
Following EE2, the ITT's mandate was
expanded to define and conduct a series of extended simulations
of ISO operations during July/August 1995. These simulations were
run in two steps, with the objectives of:
By the end of the second set of extended simulations, substantial improvements had been made in all areas of operational readiness. Because of the demands put on the operations teams during the first half of 1995, it was feared that, rather than further increasing operational readiness, a third set of extended simulations so close to the launch might have been counter- productive.
The ISO Science Operations Room at Villafranca
Off-line processing display for an ISOCAM observation, taken
after launch. The picture shows a two-dimensional image display
of flux/pixel and the corresponding three-dimensional histogram
for images collected from NGC 4038/39 ('The Antennae')
Summary of ISO Key Events Mid 1 9 9 1 Establishment of ISO core integration and Test Team Feb. 1 9 9 2 Test of the Operational Data Server Interface Apr. 1 9 9 4 EEO Test ESOC-ESTEC Jun. 1 9 9 4 EEO repeat at Villafranca Jan. 1 9 9 5 EE1 Dry-run (with simulator) Feb. 1 9 9 5 EE1 (with spacecraft) Apr. 1 9 9 5 EE2 Dry-run (with simulator) May 1 9 9 5 EE2 (with spacecraft) May 1 9 9 5 Extension of ITT terms of reference Jul. 1 9 9 5 Extended simulations phase 1 Aug. 1 9 9 5 Extended simulations phase 2 Nov. 1 9 9 5 ISO LAUNCH
In the event, the integration and validation of the ISO ground segment did indeed turn out to be a very challenging task. Despite a very tight schedule, however, and quite a few unpleasant surprises encountered along the way during the various phases of testing, the necessary validation was completed on time and within the resources allocated.
The ultimate success of the simulation campaign, namely the smooth satellite operations after launch, and the high standard of operation of the Villafranca ISO Control Centre itself since the very beginning of the mission, all bear witness to the validity of the overall approach that was adopted.