Model Based System Engineering (MBSE)
For Accelerating Software Development Cycle
Manish Patil
Sujith Annamaneni
September 2015
1
Contents
1. Abstract ............................................................................................................................ 3
2. MBSE Overview ................................................................................................................ 4
3. MBSE Development Cycle ................................................................................................ 5
4. Migration of Legacy code to MBSE ................................................................................ 10
5. MBSE Frameworks.......................................................................................................... 11
6. Benefits of MBSE Approach ........................................................................................... 14
7. Conclusion ...................................................................................................................... 16
8. References ...................................................................................................................... 16
9. About The Authors ......................................................................................................... 17
About L&T Technology Services ............................................................................................. 18
2
1. Abstract
In today’s vehicles, electronics account for about 40% of the total cost with a high
emphasis placed on software and about 200 million lines of code which are
developed in close collaboration with Service Providers, Tier 1s and OEMs. In
today’s automotive vehicle development programs, the development teams are
placed globally, increasing the geographical diversity resulting in bigger challenges
to maintain and ensure the consistent software co-ordination with different
stakeholders. The main factors behind an increase in electronic content include
increasingly stringent environmental regulations, advanced safety features, better
drivability, increasing comfort and convenience, information and entertainment
systems etc.
In the recent survey by Embedded Market Forecasters (EMF) on embedded product
development it has been revealed that impediments in the software development
process are mainly responsible for more than 80% of design delays and associated
design complications. As the embedded systems getting more complex with
increasing emphasis on software functionalities, it is becoming difficult to maintain
quality, schedule and cost within budget with the traditional approaches of
software development. In order to meet this challenge, automotive companies are
moving towards a Model Based System Engineering (MBSE) approach.
This paper demonstrates the MBSE approach which plays a crucial role in the
Software Development Life Cycle (SDLC) towards early verification and validation of
the design resulting in reduced ambiguity and delays due to design iterations along
with back-to-back testing. The benefits accrued with this approach are: reduction in
software development time, effort taken and cost.
Keywords: Model Based System Engineering (MBSE), Software Development Life
Cycle (SDLC), ISO26262, Autosar
3
2. MBSE Overview
Model Based System Engineering (MBSE) is a paradigm shift in the software
development methodology from traditional document-based specifications, design
and manual coding with more towards the executable specifications in the form of
models from which automatic code and test cases can be generated in single
development environment. The following figure shows the transition approach to
MBSE.
Figure 1: Transition to MBSE
Model-based design can be invaluable in providing early design verification and a
true executable specification. MBSE approach can be adopted during the following
scenarios:
i. Complex Functionalities: An increasing functional complexity with
continuous improvements in features results in more complexity, posing a
challenge to maintain and enhance the functionalities without disturbing the
existing working behavior.
ii. Stricter Regulations: With the advent of demand towards ensuring stricter
regulations, safety constraints (ISO26262) and higher quality, MBSE
approach supports achievement of these objectives by providing back-to-
back testing and improving quality of software early in the development
cycle.
4
iii. Reduced Development Schedule and Cost: Due to increased competition
within the automotive OEMs, the time to market for new vehicles has
reduced drastically, impacting on the overall product development life cycle
with increasing impact to reduce the schedule and cost of vehicles.
iv. Early Prototyping: Expensive or impractical to build early prototypes or
proof-of-concept development resulting in design cycles that are more
predictable and give faster results.
3. MBSE Development Cycle
MBSE approach provides a design environment that enables developers to use a
model for data analysis, visualization, testing and validation. Once the model is
developed and tested, using auto-code generation method, software for the
production embedded design is automatically generated, thereby bringing
uniformity in design and code, reducing code development time which will result in
reducing overall costs as compared to the traditional development approach.
The MBSE SDLC is outlined in below figure:
Figure 2: MBSE SDLC
5
The below figure represents a typical MBSE development cycle for capturing system
level specifications (Plant Model), control algorithms (Control Model), auto-code
generation and verification/validation methodologies.
Plant Model Control Model
Auto-code
Interface
Model-In-the-Loop Hardware-In-
(MIL) & Software-In- the-Loop (HIL)
Loop (SIL) Testing Vehicle Testing
Testing
Figure 3: MBSE Development Cycle
The MBSE workflow includes the following steps:
i. Create an executable specification consisting of algorithms, system model,
and system level verification environment
ii. Verify the system model against functional requirements using simulation
iii. Algorithm and behavioral models are optimized and refined yielding a fully
tested specification
iv. Perform floating point to fixed point conversion
v. Incorporate AUTOSAR interface information within the model for AUTOSAR
compliant auto-code generation
vi. Incorporate ISO26262 functional safety information within the model
vii. Automatically generate production software for embedded processors
6
viii. Verify the auto-code against model specifications
The below figure outlines the detailed model based development cycle with step-
by-step approach.
Figure 4: Step-by-Step MBSE Approach
The MBSE approach is being adopted for the SDLC support towards standards and
guidelines-compliance during software development stages as outlined below:
i. Standards and Guidelines Compliance
In the Model based development, Model Advisor checks supports for the
verification of the model for conditions and configuration settings of various
standards as AUTOSAR, ISO26262, IEC61508, MAAB modeling guidelines and
MISRA-C compliance. It produces a report that lists all the sub-optimal conditions
and settings that is found and suggests better model configuration settings
wherever appropriate.
ii. AUTOSAR Compliance
7
Import and export of AUTOSAR Software Component descriptions and generation of
AUTOSAR production code can be achieved in model based design. Model based
design and auto-coding tools support AUTOSAR compliance using model
configuration settings rather than AUTOSAR-specific blocks. As a result, a single
Model can be used as a reference for simulation, rapid prototyping and production
code generation in both AUTOSAR and non-AUTOSAR environments.
iii. ISO26262 Compliance
The development of safety complaint systems within the automotive industry is
characterized by demonstrating compliance with ISO26262, a standard for road
vehicle functional safety. We can generate project-specific artifacts, including
traceability matrices covering requirements, models, and generated code. Project
and product-specific artifacts can be combined to produce needful documentation
for achieving ISO 26262 certification.
iv. Verification and Validation
a. In the Model Based Design approach, executable models can be tested and
the test information can be re-used and applied later for testing code.
b. Test cases can be automatically generated or easily adopted with the
changes and coverage can be taken forward.
c. Requirements-Based Testing, Decision, Condition, Modified Condition /
Decision coverage can be covered with Model-In-the-Loop (MIL), Software-
In-the-Loop (SIL), Processor-In-the-Loop (PIL) and Hardware-In-the-Loop
(HIL) testing methods.
v. Model Reusability
Model based development approach supports the creation of library blocks which
can be developed for commonly used functions and can be included in Simulink
library toolbox for re-use into different modules or features. Modification of library
block at one place will imply the changes in all the models using these blocks.
8
vi. Retention of Legacy Code
An important feature of model-based design approach is its ability to reuse existing
pre-tested and deployed code as a part of the model. In this way, existing
applications can be updated or enhanced without having to model the complete
design. The legacy code can be encapsulated in the model using S-function
providing the appropriate interfaces for inputs and outputs mapping. Thus ensuring
the unified development environment with integration of legacy code into model
based environment for verification of the algorithm behavior.
vii. Requirement Traceability
The model can be linked with requirements in DOORS, Word or Excel format and
traceability report can be generated directly from the model.
viii. Continuous Integration and Testing
The below figure explains the automated steps adopting MBSE with continuous
integration and testing framework using Jenkins.
Figure 5: MBSE in Continuous Integration and Testing Framework
The continuous integration framework is adopted in the Model Based Design. As
soon as the developer commits the changes to a subversion repository, continous
9
integration framework process is initiated. Jenkins server detects the changes in the
subversion repository and the pre-configured Jenkins job is triggerred
automatically. During the execution of Jenkins jobs, MBSE quality gates are verified
for the following scenarios:
a. Modelling Guidelines Check: In this stage, the modelling guildelines are
verified using Matlab for MAAB compliance and the modelling guidelines
compliance report is generated.
b. Interface Check: In this stage, the inputs and outputs of the signals are
compared in the Matlab model against the data dictionary and interface
check report is generated.
c. Auto-coding Check: In this stage, the auto-code generation is invoked
and verified for the successful code generation using defined
configuration.
d. Compilation: In this stage, the auto-code compilation is carried out and
checked for any errors.
e. Unit Testing: In this stage, the unit testing in MIL (Model-in-the-Loop)
and SIL (Software-in-the-Loop) environments is carried out against the
test cases. The tests are run in Matlab and Tessy environments and
reports are generated for the coverage based on statement (C0), branch
(C1) and MCDC configuration.
During the execution of the above stages, an email will be sent to the developer on
whether passed or failed, along with a link of the execution status for each stage for
a detailed report. For the errors reported, the developer performs an analysis and
rework for the reported issues. Then the changes are re-committed to the working
copy or subversion branch which triggers the Jenkins job and the process is
repeated.
4. Migration of Legacy code to MBSE
10
The below figure outlines the overall step-by-step approach for migration of legacy
code to MBSE along with intermediate quality gates towards modelling guidelines,
MIL and SIL testing methods.
Figure 6: Flowchart for Migrating Legacy Code to MBSE
5. MBSE Frameworks
We have extensively worked for the development of automation frameworks using
Matlab scripting language for the auto-code generation and MIL / SIL testing.
i. Auto-code Generation Framework:
This framework incorporates the following aspects:
a. Configuration settings for auto-code generation
b. Customized library block checking
c. Modelling Guidelines checker
d. Code Generation utilities using Embedded Coder
e. Interface checks between model and data dictionary to ensure correct usage
of memory, boundary values and data types
11
The below figure outlines a typical Auto-code Generation Framework:
Figure 7: Auto-code Generation Framework
ii. Unit Testing Framework:
This framework incorporates the following aspects:
a. Generation of automatic test vectors for boundary values and
equivalence testing in reference with the data dictionary
b. Creation of MIL and SIL test framework with incorporation of model and
auto-code under test automatically
c. Apply test scenarios automatically and validate the results of MIL and SIL
tests
d. Generation of model and code coverage reports
The below figure outlines the unit testing framework:
12
Figure 8: Unit Testing Framework
This framework supports:
a. Detection and reporting of violations between model and code with
respect to the interface requirements
b. Identification and reporting of the precision errors between model and
code
c. Detection and report of incompatible behavior between model and code
iii. HIL Testing Environment:
Hardware-in-the-loop (HIL) simulation is a technique that is used in the
development and test of complex real-time embedded systems. HIL simulation
provides an effective platform by ensuring complexity of the plant is under the
control of the test platform. The complexity of the plant under control is
included in test and development by adding a mathematical representation of
all related dynamic systems. These mathematical representations are referred
to as the “plant simulation”. The embedded system to be tested interacts with
this plant simulation. The below figure outlines a typical HIL testing environment
being incorporated in MBSE.
13
Figure 9: HIL Testing Environment
6. Benefits of MBSE Approach
Organizations that adopt Model-Based Design or approach realize their savings
range from 20-60%, when in comparison to traditional methods. The bulk of these
savings come from better requirement analysis combined with early and continuous
testing and verification. As requirements and designs are simulated using models,
defects are uncovered much earlier in the development process.
14
Figure 8: Model-Based Design shifts defect discovery to early
development phases
Source: Measuring Return on Investment of Model-Based Design By Joy Lin,
Aerospace Industry Marketing Manager, MathWorks
Model Based System Engineering technologies including simulation modeling, rapid
prototyping, hardware-in-the-loop testing and automatic code generation offer
better design results and considerable savings to OEM owing to the advantage of
early detection through simulation.
Additionally, MBSE creates a structure for software reuse that permits established
designs to be effectively and reliably upgraded in a more simplistic and cost
effective manner. Model-Based Design for embedded software development lowers
costs by identifying defects early in the development process and reduces the total
number of latent defects thereby helping companies deliver higher quality systems
at lower costs and in lesser time frames.
15
7. Conclusion
Model-based design offers developers a distinct advantage over traditional product
development techniques. The ability to perform design validation and verification at
the onset of a project as well as the ability to test system segments using rapid
prototyping and automatic code generation (production code) provides a distinct
advantage to the developer.
From a design and project management viewpoint, the ability to validate at
different points during the design cycle enables management to more accurately
forecast duration and costs. Model-based design also brings the concept of “design
re-use” into the design cycle creating a framework for future savings.
8. References
i. Jerry Krasner, “Model-Based Design and Beyond: Solutions for Today’s
Embedded Systems Requirements”, EMBEDDED MARKET FORECASTERS,
American Technology International. January
2004. www.mathworks.com/tagteam/17878_91218v00_Krasner_Report.pdf
ii. Measuring Return on Investment of Model-Based Design By Joy Lin,
Aerospace Industry Marketing Manager,
MathWorks, https://www.mathworks.com/solutions/model-based-
design/mbd-roi-video/Measuring_ROI_of_MBD.pdf
iii. Autosar. Automotive Open System Architecture, www.autosar.org
iv. MathWorks, www.mathworks.com
v. Road vehicles — Functional safety — Part 6: Product development at the
software level https://www.iso.org/obp/ui/#iso:std:iso:26262:-6:ed-1:v1:en
16
9. About The Authors
Manish Patil
Manish Patil has 13 years of industry experience in the design, development and
validation of embedded systems majorly with automotive domain. He is currently
part of Transportation Business Unit in L&T Technology Services Limited focused on
Embedded Systems solutions in MBSE, ISO26262, Continuous Integration and
Software Development for the automotive industry.
Sujith Annamaneni
Sujith has 8+ years of industry experience in MBSE in the design, development and
validation of embedded systems majorly with automotive domain. He is currently
part of Transportation Business Unit in L&T Technology Services Limited focused on
MBSE solutions for the automotive industry.
17
About L&T Technology Services
L&T Technology Services is a wholly-owned subsidiary of Larsen & Toubro with a
focus on the Engineering Services space, partnering with a large number of Fortune
500 companies globally. We offer design and development solutions throughout the
entire product development chain across various industries such as Industrial
Products, Medical Devices, Automotive, Aerospace, Railways, Off-Highway &
Polymer, Commercial Vehicles, Telecom & Hi-Tech, and the Process Industry. The
company also offers solutions in the areas of Mechanical Engineering Services,
Embedded Systems & Engineering Application Software, Product Lifecycle
Management, Engineering Analytics, Power Electronics, and M2M and the Internet-
of-Things (IoT).
With a multi-disciplinary and multi-domain presence, we challenge ourselves every
day to help clients achieve a sustainable competitive advantage through value-
creating products, processes and services. Headquartered in India, with over 10,000
highly skilled professionals, 12 global delivery centers and operations in 35 locations
around the world, we constantly find flexible ways of working, tailored to our
assignments and customer needs.
For more information, visit us at www.lnttechservices.com
© 2015 L&T Technology Services. No part of this document may be modified, deleted or
expanded by any process or means without prior written permission from L&T Technology
Services.
18
