mil/sil/pil approach a new paradigm in model based development · bitte decken sie die schraffierte...
Post on 29-Aug-2019
0 Views
Preview:
TRANSCRIPT
Bitte decken Sie die schraffierte Fläche mit einem Bild ab.
Please cover the shaded area with a picture.
(24,4 x 11,0 cm)
MIL/SIL/PIL Approach A new paradigm in Model Based Development
www.continental-corporation.com Powertrain Division
Narayanamurthy Srinivas, Narendrakumar Panditi
Stefan Schmidt, Ralf Garrelfs
Public
Engine Systems
Agenda
8 July 2014
2Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Motivation - Model Based Development (MBD) 1
Model verified by Simulation (MvS)2
3 Case study on MIL/SIL/PIL
4
5
6
MIL/SIL/PIL Simulation results in SDA
Comparison of MIL/SIL/PIL results
Conclusion
Public
Engine Systems
Manual Model In the Loop (MIL) Model In the Loop (MIL)
Software In the Loop (SIL)
Processor In the Loop (PIL)
Specification
Design
Manual in the form of
document
Model design using MBD
MIL: Model verification
Model design using MBD
MIL: Model verification
Coding Manual coding Auto code generation
(ACG)
Auto code generation (ACG)
Code
Verification
Manual prepared test
cases to perform Unit
Testing
Tool generated test cases
to perform unit testing
Reuse MIL test cases
SIL : Software verification
PIL : Software verification on
Target processor or
equivalent instruction set
simulator
Motivation - Model Based Development (MBD)
8 July 2014
3Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Model Based Development: V- Cycle
8 July 2014
4
SOFTWARE
FLP : Floating point model
FXP : Fixed point model
MvS : Model verified by Simulation
SIL : Software In the Loop
PIL : Processor In the Loop
HIL : Hardware In the Loop
TVG : Test Vector Generation
SYSTEMRequire-
ments
Implemen-
tation
Model
(FXP)
Physical
Model
(FLP)
Automatic Code
Generation
SIL
PIL
Module
Test
Validation
HIL
Car
Integration
Test
HIL
MvS
TVG
MvS
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Model Verified by Simulation (MvS)
8 July 2014
5
Physical
model(FLP)
Implementation
Model(FXP)
C-Code(s-function in model)
C-Code(compiled for
µC- Target)
≈
=
Result
=
Result
Result
Result
ACG
MILModel
In the Loop
SIL Software
In the Loop
PIL Processor
In the Loop
Scaling
Functional
RequirementsExpected
Values
Test
cases
≈Modeling
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
MIL
Model In the Loop
SIL
Software In the Loop
PIL
Processor In the Loop
Refers to the kind of testing done
to verify the accuracy /
acceptability of a plant model or
a control system.
MIL testing means that the
model and its environment are
simulated in the modeling
framework without any physical
hardware components.
Refers to the kind of testing done
to validate the behavior of the
auto generated code used in the
controller.
The embedded software
is tested within a simulated
environment model but
without any hardware.
Refers to the kind of testing done
to validate the referenced model
by generating production code
using the model reference target.
The code is cross-compiled for
and executed on a target
processor or an equivalent
instruction set simulator.
MIL allows testing at early
stages of the development cycle.
SIL also allows to verify the code
coverage.
PIL level of testing can reveal
faults that are caused by the
target compiler or by the
processor architecture.
Definition – MIL/SIL/PIL
8 July 2014
6Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Case study on Engine Temperature function
8 July 2014
7
Test suite for calculation of load information for coolant temperature model1
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
<LOAD_TCO_MDL>
100%
V.6.5
V.7.3
V.7.2
m
u
R
x_in
x_out
IV = 0
V.7.1
C_MAF_KGH_MAX_LOAD_TCO
V.6.4
C_CRLC_LOAD_TCO_MDL
V.6.4
<LOAD_TCO_MDL>
<LV_ES>
<MAF_KGH>
Public
Engine Systems
Model In the Loop (MIL): Floating point model
8 July 2014
8
Stimuli/InputsController/Module
LevelDisplay/Verify
Simulation
project
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
||
||
||
FLP
STIMULI
MAF_KGH
V.5.8
LV_ES
V.5.8
LOAD_TCO_MDL
V.5.8
<MAF_KGH>
<LV_ES>
<LOAD_TCO_MDL>
ENTE_SIGCVTCOAIX
DISPLAY
FLP
Public
Engine Systems
Model In the Loop (MIL): Fixed point model
8 July 2014
9
Stimuli/InputsController/Module
LevelDisplay/Verify
Simulation
project
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
FXP
STIMULI
MAF_KGH
V.5.8
LV_ES
V.5.8
LOAD_TCO_MDL
V.5.8
<MAF_KGH>
<LV_ES>
<LOAD_TCO_MDL>
ENTE_SIGCVTCOAIX
DISPLAY
FXP
||
||
||
Public
Engine Systems
MvS: SDA Simulation Manager
8 July 2014
10
Deviations can be detected
and can be solved at early
stages
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Comparison results MIL - FLP/FXP (Error)
8 July 2014
11
Deviations due to
wrong scaling
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Comparison results - FLP/FXP (Corrected Case)
8 July 2014
12
Deviations are with in
the resolution
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Present Situation after MIL
8 July 2014
13
Random test cases are generated to test production code.
Execute generated test cases in the project environment.
1
2
4
3 More effort is required to prepare test cases to verify production code.
Completely different test cases are used to verify model and generated code.
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems8 July 2014
14
1Wouldn’t it be nice to reuse the MIL test cases for test of the Automatically
Generated Code ?
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Software in the Loop: SIL
8 July 2014
15
Stimuli/Inputs Sfunction block Display/Verify
Simulation
project
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
||
||
||
STIMULI
V.5.8
LOAD_TCO_MDL
V.5.8
MAF_KGH
LV_ES
LOAD_TCO_MDL
ENTE_SIGCVTCOAI
SimMode:SIL
DISPLAY
MAF_KGH
V.5.8
LV_ES
SIL SIL
Public
Engine Systems
Comparison results - MIL/SIL (wrong case)
8 July 2014
16Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Comparison results - MIL/SIL (correct case)
8 July 2014
17
No Deviation
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Processor In the Loop: PIL
8 July 2014
18
Stimuli/Inputs Sfunction block Display/Verify
Simulation
project
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
STIMULI
LOAD_TCO_MDL
V.5.8
MAF_KGH
LV_ES
LOAD_TCO_MDL
ENTE_SIGCVTCOAI
SimMode:SIL
DISPLAY
MAF_KGH
V.5.8
LV_ES
V.5.8
||
||
||
PIL PIL
Public
Engine Systems
Comparison results - SIL/PIL
8 July 2014
19Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
PIL results for different target processors - Reusability
8 July 2014
20
No Deviations
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
Public
Engine Systems
Conclusion
8 July 2014
21
Necessary test effort can be essentially minimized across simulations.1
Tests suites are portable and reusable.2
3Cost-efficient consistent testing for all phases of the development:
One test suite for all development phases (MIL, SIL, PIL).
4
5
6
Early malfunction detection.
Eases the updating of test suites for changed requirements.
Shorter development process resulting in significant time-to-market advantage.
Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs © Continental AG
top related