read - STScI
VEX MPS – supporting the
Venus Express Mission Planning Concept
Bruno Teixeira de Sousa
VEGA IT GmbH at ESOC/ESA
Robert-Bosch-Str. 5, D - 64293 Darmstadt, Germany
[email protected]
Andrea Accomazzo
ESOC/ESA
Robert-Bosch-Str. 5, D - 64293 Darmstadt, Germany
[email protected]
Ian Shaw and Robin Steel
VEGA IT GmbH
TIZ - Robert-Bosch-Str. 7 - 64293 Darmstadt, Germany
[email protected], [email protected]
Abstract
Venus Express is a deep space mission with the following
characteristics:
•
A considerable signal propagation delay on the
communication link,
•
Visibility coverage during one third of the day from the
prime ground station at Cebreros-Spain,
•
Support from additional ground stations during routine
phase only for the Radio Science experiment,
•
Capability to collect a great amount of data that needs to
be downloaded within the duration of a contact period,
•
Limited resources in terms of data link bandwidth, power
consumption during eclipses and thermal load capability
during pointing.
These factors, among others, determine that operations have to
be carried out autonomously from the Mission Timeline
mechanism implemented in the on-board Data Management
System. This in turn requires that any plan of observations for
any particular time period has to be developed and validated
well in advance.
The approach to this plan development process consists of
evolving it from an unrefined state to a more detailed planning
state. In this way we are able to freeze spacecraft resources
such as pointing, as early as possible to give the Venus
Express Mission Operations Centre (VMOC) – composed of
Flight Control (FCT) and Flight Dynamics (FDyn) teams –
enough time to evaluate and correct the observation requests
from the Venus Express Science Operation Centre (VSOC).
To achieve this, a concept that utilises multiple planning
cycles has been put in place.
The VEX MPS (Mission Planning System) uses the planning
and scheduling software kernel inherited from the ENVISAT
and Mars Express Mission Planning Systems. However, in
order to support the VEX Mission Planning concept for the
routine operations phase, a number of new functionalities have
been added to the software. A completely new software
configuration has also been implemented. These allow the
MPS to validate that the VEX mission operational rules and
constraints are respected for any particular planning cycle, and
to generate the corresponding commanding for both the
Spacecraft and Ground Stations. This paper covers the VEX
mission planning concept and the approach taken in modifying
and configuring the MPS.
VEX Planning Requirements
Venus Express was launched on the 9th of November 2005
and was inserted in orbit around Venus on the 11th of April
2006. It is carrying seven different instruments that are
operated daily around the planet. The mission is served by a
single ground station for data downlink (Cebreros – Spain);
additional stations (ESA New Norcia, Australia, and DSN
Canberra and Goldstone) are scheduled during the routine
phase for the Radio Science experiment.
The mission entered its routine mission phase in mid-May
2006 after having achieved the target operational orbit
(250x66000 km) around the planet. Payload operations are
coordinated by the Venus Express Science Operations Centre
(VSOC) – located in ESTEC, The Netherlands – which
converts scientific measurement requirements into operation
requests. These are further processed by the Venus Express
Mission Operations Centre (VMOC), located in ESOC,
Germany.
The VMOC has the ultimate responsibility to ensure that all
mission operations planned, and then actually conducted, do
not violate any operational constraints and do not exceed the
allocated resources of the space and the ground segment.
In order to guarantee smooth, safe, but still optimised science
operations a full process has been put in place between the
involved centres for the planning of mission operations.
Operational requests are processed and analysed early
enough in the planning cycle to allow time to review them
and eventually update them to match the actual capability of
the mission. The ultimate aim of the planning exercise is to
make full use of the “system”; such that the mission
objectives are achieved in accordance with the available
resources.
The planning problem can be typically seen as an
optimisation and control problem. The optimisation part is
performed at science planning level by the VSOC, whilst the
control part is executed by the VMOC. This paper deals with
the planning problem given by this second part which can be
classified in the following areas:
ƒ Plan available resources according to the science needs
and make them available for science planning (e.g.
distribution of ground station contacts has to be adequate
for requested operations)
ƒ Process incoming operation requests (e.g. include
payload operation requests in the planning)
ƒ React to incoming operation requests with additional
operations (e.g. dedicated ground station and spacecraft
configurations in the case of radio science operation
requests)
ƒ Schedule spacecraft routine operations (e.g. management
of the TTC subsystem, of the mass memory, orbit
control manoeuvres, etc.)
ƒ Verify that scheduled operations do not create any
conflict (e.g. payload status vs. spacecraft status)
ƒ Verify that scheduled operations do not violate any
constraints (e.g. no Sun illumination on the instruments)
ƒ Verify that scheduled operations are compatible with
scheduled/available resources (e.g. mass memory
overflow, or battery depletion)
ƒ Put the operations request in place (e.g. generate
commands for the spacecraft)
The complexity, criticality, and the repetitive nature of the
activities described above have driven the requirement for a
software system to be used by the VMOC personnel.
VEX Planning Concept
The planning concept is based on an iterative process
between the actors during which operations are iteratively
refined and the required level of checking is performed. The
process has been logically split into three steps:
ƒ Long Term Planning: ground segment resources are
allocated to the VSOC for science planning
ƒ Medium Term Planning: science planning is prepared and
then checked for feasibility
ƒ Short Term Planning: overall planning is further refined,
checked, and finally translated into operational commands
for the spacecraft and the ground stations.
GS usage +
Bandwidth
GS usage +
Bandwidth
LTP (6 months)
OR + PTR
checks
OR + PTR
checks
OR + PTR
checks
...
MTP 1
Launch
MTP2
STP 1
STP 2
VOI
1...
4
5
6
MTP n
STP m
STP 6
STP 3
...
STP 7
STP 4
0
...
STP 5
7
STP 8
8...
months
Figure 1 - Planning Cycles
Each planning cycle is defined by processes and interfaces
which are described hereafter.
Long Term Planning
This planning cycle covers a period of 6 months of
operations, for which the VSOC defined a Science Activity
Plan with the preferred observations for that season. Products
are delivered to VSOC 6 months prior to start of the
corresponding period, and include:
ƒ The long term orbit file (ORVF) and the orbital event
file (EVTF) from FDyn giving information about the
reference orbit to be flown.
ƒ The Downlink and Radio Science opportunities from the
available Ground Stations (this file is called FECS) and
TM Bit Rate bandwidth available (BRF) from the FCT.
Required MPS Capabilities. To support this LTP cycle the
FCT required from the Mission Planning System the
capability to:
ƒ Ingest orbit events and display a timeline of the
planning skeleton.
ƒ Process the Cebreros Ground Station Availability files,
and based on this apply specific rules to derive additional
events which determine when the Spacecraft should switch
on its transmitters and the span of the data download
windows over Cebreros contact passes.
ƒ Determine Radio Science (RSI) window opportunities
from the availability of New Norcia and DSN stations.
ƒ Ingest the Orbit Correction Manoeuvres windows as
planned by FDyn.
2
ƒ Validate that payload/subsystem mode transitions are
made in a specific order (state diagram implementation)4.
ƒ Check for incompatible payload/subsystem mode
against pointing type or event (such as illumination)5.
ƒ Derive new events from operations requests6.
ƒ Generate resource (power/data) profiles over time.
Manipulate existing profiles to generate new ones (typically
the following operations need to be performed between
profiles: addition, multiplication and integration over time)
ƒ Check for profile threshold violations
ƒ Raise warnings and display a layered timeline of events
and operations;
ƒ Profile plotting and printing capabilities.
ƒ Finally all this information is collected in an agreed
output form (the FECS mentioned above) and sent to
VSOC for detailed science planning.
Medium Term Planning
This planning cycle covers one month of operations, and
starts two months before the first operation takes place.
VSOC produces a spacecraft pointing plan, and preliminary
instruments operations requests. FDyn validates the pointing
with respect to thermal and illumination constraints and
spacecraft manoeuvring capability (wheel torque and
momentum levels). If this is feasible, FDyn then generates
the Sun aspect angles file (SAA) and medium term orbit and
attitude event file (EVTP). These are a refinement of the
EVTF but containing also attitude information such as
payload Field of View (FOV) illumination events. The FCT
checks then all products against operational rules and
constraints.
Required MPS Capabilities. To support this cycle the FCT
required the following capabilities from the Mission
Planning System:
ƒ Ingest and process any event file in a standard format.
ƒ Ingest and process of any type of operation requests1
(OR), as long as the operation exists as a sequence of
commands in the mission database.
ƒ Ingest and process the pointing file, containing
pointing type blocks spanned over time2.
ƒ Ingest the Sun Aspect Angle file (SAA) and interpolate
between each sample to produce a one second resolution
profile used to compute power generation from the Solar
Arrays.
ƒ Ingest the One Way Light Time file which allows
taking this factor into account when scheduling spacecraft
operations against ground events, or ground station
scheduling against spacecraft events.
ƒ Ingest the TM Bit Rate file to allow computation of
data dumping profiles.
ƒ Recognize particular operations as payload/subsystem
modes and allocate default power consumption/data rate
generation whenever the operation is requested3, thus
building a profile over time. Alternatively use additional
data supplied in the OR to build such resource profiles (socalled Z records).
1
POR for payloads, SOR for spacecraft generic operations
and FDR for Flight Dynamics AOCS commanding
2
Pointing types include Earth, Nadir, Slew, Inertial, Wheel
Off Loading, etc.
3
For example switching ON a payload will cause the power
consumption to raise from zero to a positive value, and an
amount of housekeeping data to be produced.
Short Term Planning
This planning cycle is performed weekly. VSOC generates
final payload operations requests with all command
parameters finalized. FDyn updates the frozen pointing file
with latest orbit prediction data and generates a new updated
SAA. FDyn also generates all the AOCS7 commanding8
corresponding to the pointing requested and creates the short
term event file (EVTV) with AOCS mode change events and
updated orbit/attitude prediction. FCT performs final check
and generates Detailed Agenda Files (DAF - command stack
for the spacecraft and schedule for the ground stations). The
spacecraft commands upload strategy for the week is decided
at this level and it is based on the total number of commands
generated, the type of operations and the workload profile.
Typically two or three uploads may be needed to avoid
overloading the on-board Mission Timeline, but the overall
concept can be pushed to have daily uploads.
4
For example payloads which require operations to be
scheduled in a particular order are checked for correctness.
5
For example certain payloads shall be OFF when
performing a Wheel Off Loading, or certain payloads shall
be OFF if the FOV is illuminated.
6
For example when receiving a request for an RSI
observation the corresponding start and end events are
derived in order to be able to command the switching of the
transmitters relative to them. When the SPICAV shutter is
commanded derived events are generated to determine the
status of the shutter and verify that it’s left closed at the end
of an observation.
7
Attitude and Orbit Control System
8
Basically all the slews and inertial or dynamic pointing
commands
3
Required MPS Capabilities. To support this cycle, in
addition to the Medium Term capabilities referred above
which are also required at this level for the final checking,
also the following capabilities are required:
ƒ To consolidate a fully checked and validated plan into
a superseding and growing Master Plan from which all the
schedules are to be generated.
ƒ To produce, from that Master Plan, sets of schedules
containing detailed commanding both for Spacecraft and
Ground Station9, for a given time range, expressed in UTC
or Orbit range (orbits start at Apocentre).
ƒ To transmit these schedules to their respective
destinations10.
ƒ To calculate total number of commands and check for
maximum commanding rate violations. The on-board time
tagged queue (called Mission Timeline or MTL)
mechanism cannot store more than 3000 commands and
execute more than 20 commands per second.
Planning Input: It is the task that retrieves and validates all
planning products sent by external parties (VSOC and FDyn)
via the FTS11.
Master Operations: It is the task that ingests and processes
event files and pointing timeline files, for a particular Orbit
Range. It also has the capability to generate new events
derived from the ingested ones (produced by FDyn, and
other external parties), based on configurable rules12, and
adding them to the Master Event Database (MPEF).
The VEX MPS
EKLOPS - The original software kernel
The existing Mars Express Mission Planning System (MEX
MPS) was used as the starting point for the Venus Express
MPS, allowing a stable system to be rapidly provided under
tight budgetary and schedule constraints. In turn, the core of
the Mars Express MPS originated from a slimmed-down
version of the ENVISAT Flight Operations Segment MPS –
with generic elements being identified as potential candidates
for reuse.
In this section, we chart the development of the software
during the Venus Express MPS implementation, and how the
underlying planning kernel has been enhanced further by this
process. This kernel, now referred to as EKLOPS (Enhanced
Kernel Library for Operational Planning Systems), also
forms the basis of several other planning developments being
undertaken on behalf of the European Space Agency.
Figure 1 – Original EKLOPS diagram
Care was taken to provide solutions which did not rely on
significant amounts of mission-specific code being written;
EKLOPS almost exclusively uses a configuration database in
order to implement mission specific functions.
EKLOPS provided already the following capabilities:
11
9
Cebreros Ground Station is currently able to receive a
schedule file containing a sequence of jobs, which almost
completely automate the station operations
10
the Mission Control System for the spacecraft schedules,
and the Ground Station Control Center (GFCC) for the GS
schedule
File Transfer System – task running in several nodes that
pushes and pulls files on user demand.
12
for example, determining the event representing the instant
that the transmitter should go ON based on the known event
of the GS starting to track the spacecraft, by subtracting an
offset and a OWLT (one way light time – to convert from
ground to spacecraft time),
4
Plan Engine: It is the heart of the system: a task to build
plans by ingesting and processing Operation Requests,
building and visualizing resource profiles, checking for
conflicts, visualizing the operations timeline, checking
boundary conditions of plans.
Scheduler: It is the task that uses the mission TC database to
convert sequences from a plan into full command stacks
(schedules) with respective parameters.
Manual Request Editor: Basic ASCII editor to write
Operations Requests. No validation or field editing
capability. This has been greatly improved for VEX.
DB test and DB editor: tasks to process XML configuration
files into the MPS configuration database, and to visualize its
contents.
Technology: The Mission Planning System has been
implemented on SUN ULTRA workstations, with C++ as
implementation language. The design has been carried out
with OOL methodologies, using UML with the Rational
Rose tool on PCs. The implementation relies on the use of
the following COTS:
ƒ POST++ and XML are used for Database Management
System;
ƒ The ILOG Views library is used for the development
of the HCI Components.
The system is delivered as libraries of re-usable components.
Especially the rule-based components of the MPS are
delivered as a library of standard conditions and actions,
which can be used to create rules.
The new software functionalities
In order to support the MPS configuration envisaged for
VEX a number of new kernel functionalities have been
implemented, namely:
ƒ The capability to process Ground Station Availability
files. These files follow a standard used both by DSN
(NASA) and ESA stations. Each file essentially indicates,
for a particular GS, the times at which the activity for a
particular pass starts and ends, and the times that tracking
will start and end. These times are converted into events
within the MPS Master Event Database (MPEF).
ƒ The capability to insert user events into the MPS
Master Event database (MPEF). There are certain
constraints that users might want to take into account when
planning, for example we know that due to a limitation in
our Solid State Mass Memory (SSMM) for specific
periodic on-board times, which occur every few months,
the storage of data has to be interrupted for a few seconds.
This functionality allows the planner to ingest the time of
this event and schedule the operation to stop and resume the
storage around it.
ƒ The capability to assign priorities to FDyn event files
for different planning cycles. That is Short Term event files
shall take precedence over Medium Term ones because
Figure 2 - Inserting user events
they contain a finer orbit determination, and Medium Term
will take precedence over Long Term, because they contain
additional data (for example contains payload illumination
events, since in this cycle (MTP), the attitude profile is
already known).
ƒ The capability to output any configurable set of events
from the database of the MPS into an event file. This was
used to output the Communication and Radio Science
windows for VSOC at long term planning (the FECS).
ƒ The capability to ingest “other data” from Operations
Requests (OR). Normally OR contain command sequences
as defined in the MIB (Mission Information Database - with
all Telemetry and Commands) for operating payloads. For
Radio Science, where there is no real physical payload, the
Science is performed using the platform transmitters. The
commanding of the transmitters by FCT and scientist
simultaneously is incompatible. Therefore the scientist uses
the operation request to pass his desired start and end of
observation, using a “dummy sequence”. The MPS when
receiving such a dummy sequence generates an event
within its database which indicates the type and timing of
the observation. The events are used for scheduling the
corresponding transmitter and GS commanding, in a
combined, and conflict free manner.
ƒ The capability to use either state based information to
build a resource profile or use additional resource
information supplied by the OR in the form of “Z record”
parameters. When requesting a particular sequence to be
scheduled the user can assign to that sequence an estimation
of power consumption (ZP record) or data rate generation
(ZD). The resource profile will then be built using these
records as markers. In their absence the default value for
that sequence/state is taken instead.
5
ƒ The capability to edit and generate Operations
Requests (OR) with an appropriate editor. Operations
requests follow an interface standard defined for the
ROS/MEX/VEX missions which define a file structure with
several editable fields. The editor uses the command
sequence database (MIB) to simplify the task of entering
sequences and to ensure their validity. It also checks that
the events used for timing exist and are valid. It formats the
output according to the standard. This is used by the FCT to
schedule additional platform commanding.
ƒ The capability to use the short code of command
sequences as the key for identifying the sequence within the
MPS database. The MEX approach was to use the
Sequence description field, since the short code was
changing with each new version. For VEX this short code is
invariant and the unique identifier. This change represented
a very deep and important change in the MPS architecture,
since the MIB database (database of sequences and
commands) had to be merged with the MPS configuration
database to allow both approaches.
ƒ The capability to have state models and automatic
mode transitions after a specified time, for elements in the
MPS database that are used to represent resource
consumption. Typically a subset of an instrument’s
command sequence is selected based on their impact on
power consumption and data rate generation. The previous
version of the MPS allowed the resource consumption to
change only at commanding of a new sequence. Therefore,
it was difficult to model behaviours like the VIRTIS
payload cooling mode: after the corresponding sequence
has been commanded a peak of power is seen for about 2
hours due to an electric motor being activated and then it
drops a significant amount when the engine stops without
any other sequence having been commanded. Also VIRTIS,
consisting of 3 independent parts (the Main Electronics and
the M and H channels), has sequences which influence the
resource behaviour of only one or possibly more of the
parts. Consequently the whole state diagram for VIRTIS
has been decoupled from the sequences, and a state model
for each of the parts has been put in place and combined to
get the total payload resource usage behaviour.
ƒ The capability to trigger warnings in the application
logging area, based on the evolution of resource profiles.
This allows the planner to visualise trends. For example the
application will report the volume of the SSMM at crossing
of various thresholds. The planner is interested in whether
the maximum volume of the SSMM has been reached, but
normally the limiting factor is actually the capability to
download the data, which is changing with the distance to
the Earth. In this case the trend of the volume becomes
more interesting, rather than whether it has reached the
maximum size.
ƒ The capability to process the Master Event Database
(MPEF, containing all the ingested FDyn and Ground
Stations events, plus corresponding derived events), in a
rule based approach. This approach was already used at the
plan processing level, that is at the level when operations
requests are processed: For each planning element
(instrument or payload or resource), separate rules exist to
ingest the OR, compute the mode models, compute
resource profiles and perform checks on the profiles and
modes. The planner can process all or only a subset of those
rules, thus building his plan in several stages, depending on
the data he has available for processing. The planner also
saves time by only processing individual rules if his
interested is in only in a particular part of the plan. The
same rule based approach is used for the task of preparing
the plan background, which involves the ingestion of
external events and the derivation of new ones – the Master
Ops. That is the planner can now choose which activities to
perform to build his MPEF by selecting appropriate rules.
Figure 3 - Master Ops - Rule based approach
ƒ The capability to generate schedule files containing the
commands to drive the ESA stations during a VEX pass.
These commands are timed against the information
obtained from the above mentioned GS Availability files.
Currently the Cebreros Ground Station is automatically
driven by the schedules generated from the VEX MPS
without the need for operator intervention. The passes are
initialized, the tracking and uplink are started and
terminated and the pass finalized autonomously. In the near
future the Radio Science passes of VEX over New Norcia
and dedicated schedule of events for DSN (NASA stations)
will also be produced by the MPS.
ƒ The capability to generate schedules for a particular
time range expressed in UTC. In the previous version of the
MPS it was only possible to generate schedules for Orbit
6
ranges. With this new functionality the start and end times
of the schedule can be carefully chosen not to split any
operations. The orbit boundaries, being the Apocenter,
often coincides with on-going observation or pass activities
which would be split into different DAF. If one of the DAF
for some reason is not uploaded13 then the spacecraft could
be left in an inconsistent state.
The “Operational Rules and Constraints” and
the VEX MPS configuration
A set of routine mission operational rules and constraints has
been identified in a document (the ORCD) by the FCT based
on the Spacecraft User Manual as laid down by the
manufacturer. These rules and constraints drive the checks
and modelling that the MPS makes in order to validate a
particular plan of operations. Therefore this document and
the MPS configuration are very close coupled. They are in
fact one and the same baseline. They are maintained together
under configuration control, and any changes to the
document are reflected in the configuration and vice-versa.
Master Event Database. The first part of the document
concerns how the external environment (orbit, attitude and
GS activity) is represented within the MPS.
ƒ The inputs that have to be ingested are identified, such
as: the several FDyn event files14; the FDyn pointing
timeline (FTLV); and the GS availability files (ESAF and
DSN SAF).
ƒ The valid types of spacecraft pointing attitude are also
identified, such as EARTH, NADIR, INERTIAL, etc.;
ƒ It is described how new events should be derived, such
as: the instant the transmitter should be switched ON or
OFF, relative to the time the GS will start or end tracking
respectively; the instant at which the data downlink should
start or end relative to the time the transmitter goes ON or
OFF; the Radio Science windows relative to the start and
end of NNO or Canberra passes.
ƒ The subset of events that needs to be output at Long
Term cycle for the VSOC is defined.
This is the data that is processed by the Master Operations
task within the MPS.
Element modelling and conflict checking. A second part
addresses how the different elements of the spacecraft, which
are relevant15 for planning, are modelled, based on the
commanding profile as received from Operations Requests
(OR), in order to determine conflicts. The transmitter status
has significant impact on power; the type of pointing will
influence the thermal power consumption16; each payload
will have a power consumption and data rate generation
profile depending on the operating mode. The conflict checks
are also defined, such as:
ƒ Some payloads shall be OFF when the spacecraft is
performing a Wheel Offloading or an Orbit Correction
Manoeuvre.
ƒ Some payloads shall be OFF or their shutters closed
when direct sun illumination occurs (event identified by
Flight Dynamics based on the attitude profile)
Figure 4 - Payload illumination conflict detection
ƒ Some payloads shall be OFF at the end of an
observation.
ƒ Some payloads shall have thermal control enabled
prior to an observation.
ƒ Some payloads shall be commanded according to
particular command sequence patterns.
ƒ The High Speed Link (bus that gives some payloads
direct access to the SSMM) shall not be active when the
SSMM is being down linked to Earth.
ƒ The spacecraft shall be in EARTH pointing mode
during a communications window
13
For example GS unavailable due to failure
EVTF, EVTP and EVTV for long, medium and short term
cycles respectively as mentioned before
15
The relevant elements are those which affect resource
usage, such as power consumption and data rate generation.
14
16
Pointing with illumination on radiating faces will cause the
spacecraft to warm-up and the heater consumption to
decrease, and vice-versa.
7
ƒ The spacecraft shall only switch on the transmitter in a
cooling attitude17.
ƒ RSI observations shall only be performed within RSI
passes with NNO or Canberra.
The processing of element modelling and conflict checking
is realized within the MPS Planning Engine task.
and exceeded limits. An exceeded limit causes the
operation being planned to be not accepted. The spacecraft
will detect a battery under charge condition and react with
Safe Mode. Ageing factor of the battery and Battery
Regulator efficiency are also taken into account.
Resource Management. A third part concerns the resource
usage profiles generated and the checks performed on them:
ƒ The total rate profile on the spacecraft data bus
(OBDH) is calculated and checked against critical and
excessive thresholds. Exceeding traffic on the bus will lead
to buffering of data and underperformance of the data
handling subsystem.
ƒ The amount of data in each science file allocated to a
payload, in the memory (SSMM), is calculated and checked
against overflow. The MPS will raise the warning it is up to
the scientists to accept the risk or modify their observation
profile.
ƒ The total SSMM volume is calculated by integrating
the total platform generated data rate (OBDH plus High
Speed link) subtracted from the available bandwidth for
downlink during passes over time and checked against the
capability to downlink within the next pass opportunities.
Once more it is the scientist’s call to accept the risk of
eventual data loss; once the SSMM overflows, old data
starts to be overwritten.
ƒ
The power generated by the solar arrays is calculated
as a function of the angle of the sun to the surface of the
panels over time, and of the distance of Venus to the Sun,
and masked by periods of eclipse. Power regulator
efficiency is also taken into account.
ƒ The total power consumption profile of the platform is
calculated by adding the average consumption of the
permanently ON instruments18, with the Transmitter and
Payload consumption (as a function of their modes), with
the Thermal Consumption (as a function of pointing) and
taking also into account harness losses; and then checked
against the maximum power delivery capability of the
Power Distribution Unit (PDU). No operations scenario
where this threshold is exceeded can be accepted. It would
result in a main bus under voltage condition that would
bring the spacecraft to Safe Mode.
ƒ The batteries power flow profile is calculated taking
into account that the batteries will supply (discharge) the
difference between the PDU demand and the solar array
available power, and when discharged with enough solar
power being produced will charge at a fixed rate. The
Depth of Discharge (DoD) results from the integration of
the charge and discharge rate, and is checked for critical
Figure 5 - Power consumption profile
Likewise, the above resource management is performed by
the Plan Engine task within the MPS.
17
18
No sun on radiating faces
Such as CDMU, SSMM, AOCS, etc.
Scheduling. Finally the last part concerns the generation of
commands for the spacecraft, which is performed by the
Scheduler task, which expands the requested sequences into
stacks of time-tagged commands, with their respective
parameters, using the mission TC database for the
conversion. Here the constraint that the Mission Timeline
mechanism on-board is not be able to process more than 20
commands per second is checked. An additional constraint is
the maximum capacity of the Mission Timeline (mechanism
that executes the commands - MTL) which can hold up to
3000 commands. Currently the MPS has no memory or
knowledge of the content of the on-board command queue
and of the uplink strategy, and cannot therefore check for this
constraint. This will be an area for future development. The
MPS does report on the size of each generated command
file19 allowing the planner to decide on the uplink strategy
according to the current capacity of the MTL.
In order to address completely the MPS configuration the
ORCD also mentions the application Timeline display
19
Also known as DAF – Detailed Agenda File
8
configuration, and the default states for all payload modes,
and resources.
All the above mentioned modelling is implemented in the
MPS using so-called (internally) “rules”. These “rules” are
defined in XML configuration files which represent the
interface with the application kernel, instantiating calls to
kernel functions20 and passing the necessary parameters so
that the specified models are achieved. The files are later
converted into a database that the application is able to
access faster. Through the MPS application MMI one can
activate and de-activate the rules that are to be processed, but
one cannot modify the rule itself, other than by editing the
XML configuration files, and re-generating the database. In
the future this is an area where more development is
welcome, such as flexible editors that allow the users to
define their problem domain and models and checks and then
having them directly published for use by the application.
The MPS validation and the end to end testing
Three formal deliveries were made prior to entering
operations and additional intermediate incremental deliveries
with patches for particular problems were also delivered.
A formal system integration test plan was developed and
executed and two end to end tests have been performed
involving all the interfaces21, allowing the MPS to be
extensively exercised prior to entering operations.
The end to end testing has proven to be the most profitable
and rewarding method, since emulation of a near-real-time
operations scenario introduced the heaviest load on the MPS,
allowing detection of problems that otherwise would not
have been detected with functional testing. It also exercised
the interfaces allowing the different systems involved to
align and correct deficiencies.
Approximately 150 Software Problem Reports were raised
and fixed by the developers, and approximately 100
configuration or planning products changes were raised
internally. All system and integration level testing was
performed by the FCT, and development management shared
with the Data Systems division. Software problems were
managed with a web-based tool from the Data Systems
Division (e-log). Internal changes were managed with the
quality supported Anomaly Report Tool (ARTS) and
products maintained under configuration control with the
Clear Case repository.
20
These functions include: function to ingest and parse a
particular type of file; function to check a mode against
another and give warning; to sum a number of profiles; to
multiply two profiles; to integrate a profile over time; derive
an event from another with offset; and many others.
21
VSOC – VEX Science Operations Center; and Flight
Dynamics.
The Future needs
The current system is a significant evolution from the
systems already used in ESOC for other missions in the
sense that it encompasses several functions which are
normally performed outside the classical Mission Planning
System. Further to this the system has been extended not
only to cover the spacecraft activities but also the ground
station activities. Nevertheless there are some areas where
the system performance needs to be improved or new
functionalities need to be added.
New functionalities required in the immediate
future
Among the functionalities and configuration updates
required for the immediate future one can envisage the
following:
ƒ New conflict checks and updated models. Since the
spacecraft has been operated in a routine manner we have
come to understand some constraints differently, this and
the need to increase flexibility in constraints that are
currently too strict, created the need to re-consider some of
the current implementations. In this category fall the checks
of Transponder usage outside of ground contact
configuration, the modelling of packet stores, and the
modelling of some payloads operations.
ƒ NNO RSI passes scheduling. As for Cebreros, the
NNO ground station will soon be ready to receive schedules
for automatic commanding.
ƒ DSN station key file generation. In the same way the
DSN stations require a dedicated file to schedule their
activities during RSI observations.
ƒ Automatic synchronized Spacecraft/Ground Station Bit
Rate changes. This functionality will allow setting of the
Bit rate in ground station commands according to the
current spacecraft settings.
Reporting Capability
At the reporting level a number of capabilities are lacking
which will be addressed, such as:
ƒ Generation of planning session reports and meaningful
summaries. This should provide evidence of the activities
performed by the planner and allow the user to trace
eventual deficiencies in the planning process that could
result in the execution of incorrect operations (e.g.. missing
operations).
ƒ Generation of Pass Sheets, to instruct the SPACONs22
on what is to be uploaded and provide pass-relevant timing
information (i.e. begin and end of tracking, carrier up and
22
Spacecraft Controllers – The person that monitors and
commands the spacecraft during a ground contact period.
9
down, when to expect eclipses and star-tracker occultation
events, etc.)
ƒ Exporting of resource profiles to worksheet format to
be manipulated and analysed by external tools (ex. Excel).
ƒ Displaying events and activities in a chronological list
format for quick review of timeline.
Schedule Generation management
Currently the MPS has no memory or record of what has
actually been generated/uploaded/commanded to the
spacecraft. In this area it could be envisaged that in the future
the MPS would be able to: keep track of deployed commands
to avoid repetitions or overlaps; and advise the planner on a
generation or upload strategy for commands, taking also into
account the command storage capability on-board. Linked to
this is also the capability to recognize that the removal of an
operation from the plan implies the deletion of the respective
commands in case they are already on-board.
The VEX MPS in numbers
Some statistics about the operational usage of this system at
the time of the final update of this paper:
- 7 Medium Term cycles fully planned with several
conflicts early detected and addressed;
- 20 Short Term cycles fully planned with more than
80.000 commands generated and uploaded to the
spacecraft and 2.000 commands for the ground
station;
- An average 4/5 hours required by the FCT for
planning each week of operations (STP), and 2/3
hours for each monthly planning session (MTP);
- Approximately 220 Gbits of science data recorded
and downloaded from Venus to the ground over
approximate 1400 hours of contact.
- One observation missed out of approximate 600
commanded, due to human error in planning
process. No observation missed related to an MPS
fault.
- 0 Spacecraft Safe Modes caused by planning errors.
- 0 dedicated planners in the VEX FCT. The task is
shared among all the 8 engineers and analyst on a
shift basis, and represents 5% of each one’s total
amount of work (approximately 10 hours every 5
weeks are dedicated to planning activities).
This represented a major achievement and mile-stone for the
Mission Operations of VEX. The VEX Flight Operations
Plan was also ready to support the planning activities as
envisaged in the planning concept with the usage of the
MPS, and all the members of VEX FCT have been trained
for using the tool and accomplishing with success the
operations planning.
The MPS has shown an advanced level of maturity and
stability as a result of the long period of testing and finetuning. The plan for the first week was produced by one
person within one working day, and a faultless schedule was
generated, proving the efficiency and effectiveness of the
tool. Work is still on-going to improve its performance;
currently some of the tasks that the planner is required to
perform with the tool can be accelerated (i.e. starting up of
applications, creation and processing of plans, saving and
committing plans). New functionality is already envisaged to
upgrade the tool to provide a greater level of flexibility and
planning capability.
As the FCT gathers experience, the planning concept will
also evolve to better correspond to the missions specifics and
the MPS will evolve with it.
References
[1] I.Shaw, M.Niézette et al., Mission Planning for Europe's
latest earth observer, Proc. Int. Sym. Ground Systems for
Space Mission Operations. SPACEOPS 2000, Toulouse,
June 2000.
[2] I.Shaw, M.Niézette et al., A new world for the Generic
Mission Planning System, Proc. Int. Sym. Ground Systems
for Space Mission Operations. SPACEOPS 2002, Houston,
Texas, October 2002
[3] A.Accomazzo. 2005. VENUS EXPRESS Mission
Planning Concept, ESA internal document.
[4] B.Sousa, I.Dauvin, et al. 2006. VENUS EXPRESS
Mission Planning System Operations Rules and Constraints
Document, ESA internal document.
[5] B.Sousa, A.Accomazzo, 2005. VENUS EXPRESS MPS
User Requirements Document, ESA internal document.
[6] I.Shaw, R. Steel, et al. 2006. VENUS EXPRESS Mission
Planning System Software User Manual, ESA internal
document
Conclusion
The VEX MPS successfully entered operations on the
beginning of May 2006. In the first week of routine
operations in Venus, VEX was already being guided with the
stacks of commands generated and validated by the MPS.
10
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement