ESOC Simulators Past, Present and Future Nuno Sebastião (OPS-GIC), Vemund Reggestad (OPS-GDA), David Verrier (OPS-GDS)

Outline of this Presentation The way we got here Current Operational Simulators Simulation Model Portability Standard (SMI/SMP2/ECSS) Infrastructure Current and planned activates Cooperation with Partners and ESTEC The Future of Operational Simulator Development What Technology can do Summary

The way we got here

Current Operational Simulators

Simulation Model Portability Standard (SMI/SMP2/ECSS)

Infrastructure

Current and planned activates

Cooperation with Partners and ESTEC

The Future of Operational Simulator Development

What Technology can do

Summary

Future of ESOC Simulators The way we got here

The way we got here

Simulator development evolution Old Spacecraft Little OBSW Processor not running, TM still generated easy to generate fixed format telemetry TM and TC hard-wired Few changes in AIT New spacecraft Lots of OBSW, FPGA etc.. If processor not running, no telemetry generated Packet telemetry and telecommands Defined late, defined in software Consequently, Many late changes

Old Spacecraft

Little OBSW

Processor not running, TM still generated

easy to generate fixed format telemetry

TM and TC hard-wired

Few changes in AIT

New spacecraft

Lots of OBSW, FPGA etc..

If processor not running, no telemetry generated

Packet telemetry and telecommands

Defined late, defined in software

Consequently, Many late changes

Operational Simulator Costs Cost have decreased. Previously Large cost overruns. Up to 100%

Cost have decreased.

Previously Large cost overruns.

Up to 100%

Cost Drivers Changes after Kick-Off No on-board standardization. Length of development Changes & delays on the spacecraft lead to changes and costs on the simulator Starting too early Number of emulators Starting from scratch

Changes after Kick-Off

No on-board standardization.

Length of development

Changes & delays on the spacecraft lead to changes and costs on the simulator

Starting too early

Number of emulators

Starting from scratch

Future of ESOC Simulators Reducing costs (Current measures)

Reducing costs

(Current measures)

Avoiding concurrent development: sim & MCS Historically: The simulator used as data source/sink during MCS development and testing. Parallel development of Simulator and MCS  Difficult to isolate problems. Solution: Develop data source/sink infrastructure configurable via database to be used for MCS validation. (i.e. GSTVi) Models based on PUS standard instead of S/C documentation. Second advantage: Permits delay to simulator development

Historically:

The simulator used as data source/sink during MCS development and testing.

Parallel development of Simulator and MCS  Difficult to isolate problems.

Solution:

Develop data source/sink infrastructure configurable via database to be used for MCS validation. (i.e. GSTVi)

Models based on PUS standard instead of S/C documentation.

Second advantage:

Permits delay to simulator development

Avoid concurrent development: sim & spacecraft Historically: Dependency between simulator and mission schedules for the ICDs, OBSW, SDB Provision of the OBSW typically end up on the critical path for the simulator Impossible to perform end-to-end test without OBSW Solution: Delay K/O of Simulator. Use GETS instead of OBSW for Simulator integration. GETS: Generic Onboard software. Better tools to isolate SDB.

Historically:

Dependency between simulator and mission schedules for the ICDs, OBSW, SDB

Provision of the OBSW typically end up on the critical path for the simulator

Impossible to perform end-to-end test without OBSW

Solution:

Delay K/O of Simulator.

Use GETS instead of OBSW for Simulator integration.

GETS: Generic Onboard software.

Better tools to isolate SDB.

Controlling development cost Measures taken: No OBSW in first delivery Integration testing via limited functional model of OBSW. Late kick-off Provision for 10 versions of flight software in FFP Full responsibility by contractor: No excuses No recourse No CCNs driven by developer Example: HP only 5% increase despite 18 month delay in launch

Measures taken:

No OBSW in first delivery

Integration testing via limited functional model of OBSW.

Late kick-off

Provision for 10 versions of flight software in FFP

Full responsibility by contractor:

No excuses

No recourse

No CCNs driven by developer

Example: HP only 5% increase despite 18 month delay in launch

Simulation Model Portability 2 Standard Outline SMP2 Objectives CCB Members Status

SMP2 Objectives

CCB Members

Status

Simulation Model Portability Standard - Objectives Enable the reuse of simulation models between: different project phases different projects different simulation platforms Reduce cost of simulator development Use modern software technologies Improve support for model reuse Improve model integration support – code generation Support integration of system engineering data Support of dynamic simulations Support for meta-data (Units, Limits) Make use of open standards (XML, UML) Decrease sensitivity to platform change (C++, Java, Windows, Linux)

Enable the reuse of simulation models between:

different project phases

different projects

different simulation platforms

Reduce cost of simulator development

Use modern software technologies

Improve support for model reuse

Improve model integration support – code generation

Support integration of system engineering data

Support of dynamic simulations

Support for meta-data (Units, Limits)

Make use of open standards (XML, UML)

Decrease sensitivity to platform change (C++, Java, Windows, Linux)

SMP Evolution 1999 – Sees the birth of SMP1 SMP1 has been applied to a number of projects: Generic Project Test Bed (GPTB) at ESTEC Rosetta/Mars Express/Venus Express/Cryosat/GOCE/HP Operational Spacecraft Simulators Galileo System Simulator Facility for ESTEC However, a number of issues identified: No support for OO/Interfaces/Components i.e. modern methodologies to support software reuse. Large amount of “glue” code has to be hand written to support the integration i.e. effort required and error prone. Does not support self descriptive models i.e. model meta-data. Platform specific - only one “platform” supported (C)

1999 – Sees the birth of SMP1

SMP1 has been applied to a number of projects:

Generic Project Test Bed (GPTB) at ESTEC

Rosetta/Mars Express/Venus Express/Cryosat/GOCE/HP Operational Spacecraft Simulators

Galileo System Simulator Facility for ESTEC

However, a number of issues identified:

No support for OO/Interfaces/Components i.e. modern methodologies to support software reuse.

Large amount of “glue” code has to be hand written to support the integration i.e. effort required and error prone.

Does not support self descriptive models i.e. model meta-data.

Platform specific - only one “platform” supported (C)

SMP Evolution

CCB Members

SMP2 - Status ECSS-E-40-07 Standard being finalised for public review (Oct 2008) Fully supported by Simsat 4 (Version 1.2) Partially supported by Eurosim, Basiles, SimTG Industrial validation conducted by: CNES (Prime) Spacebel Thales Alenia Space EADS Astrium Ellidiss “ We have developed a initial CMM implementation inspired by Simulation Model Definition Language (SMDL) of Simulation Model Portability 2 (SMP2) standards ” School of Information System and Management National University of Defense Technology, China

ECSS-E-40-07 Standard being finalised for public review (Oct 2008)

Fully supported by Simsat 4 (Version 1.2)

Partially supported by Eurosim, Basiles, SimTG

Industrial validation conducted by:

CNES (Prime)

Spacebel

Thales Alenia Space

EADS Astrium

Ellidiss

SMP2 – Industrial Validation Results Objectives Validation of SMP2 in an industrial process SMP2 tools support aspects: design, development, integration and code generation SIMSAT4 MIE (ESOC), STOOD Distributed development (multi-sites) Source and binary code sharing Integration on multiple simulator infrastructures SIMSAT4 (ESOC), BASILES (CNES), SimTG (Astrium) Simulation implemented and run successfully on SIMSAT4, BASILES SMP2 is a valid approach for development since it allows focusing on engineering rather than on infrastructure Development: real benefit (SMP2 serves as a common language among all development parties) Lack of a graphical visualization of models and Assemblies Tools need further improvements for industry adoption.

Objectives

Validation of SMP2 in an industrial process

SMP2 tools support aspects: design, development, integration and code generation

SIMSAT4 MIE (ESOC), STOOD

Distributed development (multi-sites)

Source and binary code sharing

Integration on multiple simulator infrastructures

SIMSAT4 (ESOC), BASILES (CNES), SimTG (Astrium)

Simulation implemented and run successfully on SIMSAT4, BASILES

SMP2 is a valid approach for development since it allows focusing on engineering rather than on infrastructure

Development: real benefit (SMP2 serves as a common language among all development parties)

Lack of a graphical visualization of models and Assemblies

Tools need further improvements for industry adoption.

Future of ESOC Simulators Simulator infrastructure: Paving the way to the future.

Simulator

infrastructure:

Paving the way

to the future.

Infrastructure Evolution

SIMSAT 4 SIMSAT 4 SMP2 Support MMI Migration to Eclipse Community Requests

SIMSAT – Supporting the Full Simulation Life cycle The Big picture

User-Requested Improvements Schedule Analyzer Parameter Recording Logger improvements (Filtering, Search, User Comments) Model Failure Parameter Forcing Parameter Limits Conditional and Scheduled Commanding Simulation Tree Search

Schedule Analyzer

Parameter Recording

Logger improvements (Filtering, Search, User Comments)

Model Failure

Parameter Forcing

Parameter Limits

Conditional and Scheduled Commanding

Simulation Tree Search

Universal Modelling Framework IDE like (Eclipse based) environment for the development of SMP2 based simulators. Supporting the full life cycle of Simulator development from design of models, code generation (including code merge), model execution and debugging. Will address most of the concerns expressed in the SMP2 industrial Validation study First version available Q2 this year.

IDE like (Eclipse based) environment for the development of SMP2 based simulators. Supporting the full life cycle of Simulator development from design of models, code generation (including code merge), model execution and debugging.

Will address most of the concerns expressed in the SMP2 industrial Validation study

First version available Q2 this year.

MCS NIS TMTCS IFMS/TCDS EGSE/FEE Spacecraft TM/TC Spacecraft Model MCS Model NIS Model TIF EGSE / FEE I/F Ground System under Test GSTVi Component Space/Ground I/F SGM MCS DIF IMBU (H/W) IMBU I/F AIV System Real Systems Simulated Systems The big picture Ground Segment Test and Validation Infrastructure Available to missions Q2 2008

MCS NIS TMTCS IFMS/TCDS EGSE/FEE Spacecraft TM/TC Spacecraft Model MCS Model NIS Model TIF EGSE / FEE I/F Ground System under Test GSTVi Component Space/Ground I/F MCS DIF IMBU (H/W) IMBU I/F AIV System Real Systems Simulated Systems SLE Ground Model SGM Very fruitful collaboration with the Aeolus team for the validation of the model.

Porting of Generic Models to SMP2 Positioning and Environment Model (PEM) Spacecraft Dynamics Simulation (SIMDYN) Spacecraft Thermal Network (TNET) Spacecraft Electrical Network (SENSE) Spacecraft TM/TC Toolkit (SIMPACK) First delivery accepted by ESOC. Negligible delay (One week) over a project life of 9 months. This might tells us something about the efficiency of SMP2

Positioning and Environment Model (PEM)

Spacecraft Dynamics Simulation (SIMDYN)

Spacecraft Thermal Network (TNET)

Spacecraft Electrical Network (SENSE)

Spacecraft TM/TC Toolkit (SIMPACK)

First delivery accepted by ESOC.

Negligible delay (One week) over a project life of 9 months.

This might tells us something about the efficiency of SMP2

2008 Simulator Activities Spacecraft Simulator Reference Architecture – Kick off Q1 EGOS-MF and SIMSAT MIE merge into a single product supporting the SMP-2 models development (UMF) - Q2 SIMSAT 4 Upgrades (migration to EUD, Scheduler Improvements, Win support?, 3D Adapter) – Kick Off – Q3 Support OBSW White-Box testing – Kick off Q3 Ground Station Simulator (GSTP study): Simulate the M&C Aspects of the Elements that surround the STC2 Used in the training of Operators without the real Station.

Spacecraft Simulator Reference Architecture – Kick off Q1

EGOS-MF and SIMSAT MIE merge into a single product supporting the SMP-2 models development (UMF) - Q2

SIMSAT 4 Upgrades (migration to EUD, Scheduler Improvements, Win support?, 3D Adapter) – Kick Off – Q3

Support OBSW White-Box testing – Kick off Q3

Ground Station Simulator (GSTP study):

Simulate the M&C Aspects of the Elements that surround the STC2

Used in the training of Operators without the real Station.

OBSW White Box More user friendly troubleshooting of On Board Software in the Simulator View and control the details of software execution in a "symbolic debugger" mode. Capability to execute the software in a realistic "mission environment“ ` Simulator Kernel Emulator S/C Model MCS/NCTRS ` WHITE-BOX Kernel Ground Models Simsat MMI WHITE-BOX MMIPlugin

More user friendly troubleshooting of On Board Software in the Simulator

View and control the details of software execution in a "symbolic debugger" mode.

Capability to execute the software in a realistic "mission environment“

Future of ESOC Simulators Cooperation with ESTEC (TEC-SW)

Cooperation

with ESTEC

(TEC-SW)

Converging SIMSAT and SIMVIZ ? Linux MIE SIMSAT 4 SIMULUS SMP2 Ground GENM Windows Designer SIMSAT 2 SIMVIZ SMP2 COM/Excel Models Simsat validation under windows. One simulation platform for ESA

Software Emulators Current status Novell is no longer supporting ADA (ESOC Emulator core is done in ADA) Ways forward Develop a new Emulator based on Qemu Use TSIM as the baseline for all simulation activities?

Current status

Novell is no longer supporting ADA (ESOC Emulator core is done in ADA)

Ways forward

Develop a new Emulator based on Qemu

Use TSIM as the baseline for all simulation activities?

Reference Simulator Architecture Phase A Horizontal Approach Rosetta Aeolus Vertical Approach Mars Express Phase B Phase C Phase D TEC-SW OPS-GI/GD

Common Library of Models A set of Simulator models that: Includes both generic and specific models Would be used in the Reference S/C Could be used in isolation Serve the purposes of TEC-SW and OPS

A set of Simulator models that:

Includes both generic and specific models

Would be used in the Reference S/C

Could be used in isolation

Serve the purposes of TEC-SW and OPS

Future of ESOC Simulators Cooperation with partners

Cooperation

with partners

The Industry A number of simulation platforms are available: Simsat (ESA/ESOC) Eurosim (ESA/NLR) Basiles (CNES) EcoSimPro SimTG (Astrium) etc Problem: too many platforms, so: too much effort spent on infrastructure no model share/reuse between platforms.

A number of simulation platforms are available:

Simsat (ESA/ESOC)

Eurosim (ESA/NLR)

Basiles (CNES)

EcoSimPro

SimTG (Astrium)

etc

Problem: too many platforms, so:

too much effort spent on infrastructure

no model share/reuse between platforms.

Cooperation with partners SMP2 levels the playing field. Enables reuse of models across simulation kernels UMF Modelling Framework A common Modelling environment to be used allowing proprietary/specific runtime environments The de facto tool for the modeling and assembling of SMP2 elements. Released under a open source license so that all interested parties can contribute to its development (coordinated by ESA). Foster the community development around SMP2

SMP2 levels the playing field.

Enables reuse of models across simulation kernels

UMF Modelling Framework

A common Modelling environment to be used allowing proprietary/specific runtime environments

The de facto tool for the modeling and assembling of SMP2 elements.

Released under a open source license so that all interested parties can contribute to its development (coordinated by ESA).

Foster the community development around SMP2

Future of ESOC Simulators Towards model integration

Towards model integration

Reference Simulator Architecture. Why do we need one? Current situation: No clear interface between models Difficult to isolate models for reuse Tight coupling SDB and models. Reuse by cut&paste. Objectives: Definition of clear interfaces between the different elements in a Spacecraft. Improved reusability at model level. (towards plug&play)

Current situation:

No clear interface between models

Difficult to isolate models for reuse

Tight coupling SDB and models.

Reuse by cut&paste.

Objectives:

Definition of clear interfaces between the different elements in a Spacecraft.

Improved reusability at model level. (towards plug&play)

Improvement Step 1: GMPort Generic Models interfaces clearly defined Allow plug-and-play update of generic models. First part of infrastructure related to SDB handling added.

Generic Models interfaces clearly defined

Allow plug-and-play update of generic models.

First part of infrastructure related to SDB handling added.

Reference Architecture Establish a suitable breakdown of simulators into models. Standardize on common generic interfaces between the models. Clear cut between TM/TC and engineering parameters Use the SMP2 assembly mechanism to allow different usage scenario

Establish a suitable breakdown of simulators into models.

Standardize on common generic interfaces between the models.

Clear cut between TM/TC and engineering parameters

Use the SMP2 assembly mechanism to allow different usage scenario

Future of ESOC Simulators The Future of Operational Simulator Development

The Future of Operational Simulator Development

Model Library Library of Models to include: All models developed at ESOC. Models developed at ESTEC (*) Phase-A models and SVF models (*). All models to be easily available for future developments (*) if developed according to SMP-2 and souce-code is provided.

Library of Models to include:

All models developed at ESOC.

Models developed at ESTEC (*)

Phase-A models and SVF models (*).

All models to be easily available for future developments

Step 1: Simulator Assembly If developed according to a Reference Architecture, this is feasible:

Step 2: Customization TX_A Monitoring: Bit Rate Mode Coherent Mode … Actions: Select Hi_bit_rate_mode Select Low_bit_rate_mode … Configuration: Polarization = [left, right] … SDB TM: XYZ1000 XYZ1001 … SDB TC: XYZ0001 XYZ0002 …

TX_A

Monitoring:

Bit Rate Mode

Coherent Mode

…

Actions:

Select Hi_bit_rate_mode

Select Low_bit_rate_mode

…

Configuration:

Polarization = [left, right]

…

SDB TM:

XYZ1000

XYZ1001

…

SDB TC:

XYZ0001

XYZ0002

…

Step 3: Runtime

Future of ESOC Simulators Summary

Summary

The Way Forward Start with GSTVi for MCS development and testing Consequence: Sim development can start later. Use the Reference Architecture to base the development of Operational Sims. Use GETS in D1 (replace OBSW) Shorter development cycles through reuse of models and reference architecture.

Start with GSTVi for MCS development and testing

Consequence: Sim development can start later.

Use the Reference Architecture to base the development of Operational Sims.

Use GETS in D1 (replace OBSW)

Shorter development cycles through reuse of models and reference architecture.

Future of ESOC Simulators Summary ESA is a central player in the European Simulation Domain Cooperation between ESOC and TEC-SW Building the future with: SMP2 – Improved development process Reference Architecture – More reuse Better infrastructure – Simsat 4, GSTVi, RefA More commonality – Library of Models

Summary

ESA is a central player in the European Simulation Domain

Cooperation between ESOC and TEC-SW

Building the future with:

SMP2 – Improved development process

Reference Architecture – More reuse

Better infrastructure – Simsat 4, GSTVi, RefA

More commonality – Library of Models

Thank you for your attention ! Let’s build the future Let’s continue to build the future

Let’s continue to build the future