Mil-Sil-Pil A New Paradigm in Model-Based Development

1 Bitte decken Sie die schraffierte Fl che 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 Narayanamurthy Srinivas, Narendrakumar Panditi Stefan Schmidt, Ralf Garrelfs Powertrain Division Agenda 1 Motivation - Model Based Development (MBD). 2 Model verified by Simulation (MvS). 3 Case study on MIL/SIL/PIL. 4 MIL/SIL/PIL Simulation results in SDA. 5 Comparison of MIL/SIL/PIL results 6 Conclusion 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 2. Motivation - Model Based Development (MBD). manual Model In the Loop (MIL) Model In the Loop (MIL). Software In the Loop (SIL). Processor In the Loop (PIL). Specification manual in the form of Model design using MBD Model design using MBD.

2 Design document MIL: Model verification MIL: Model verification Coding manual coding Auto code generation Auto code generation (ACG). (ACG). Code manual prepared test Tool generated test cases Reuse MIL test cases Verification cases to perform Unit to perform unit testing Testing SIL : Software verification PIL : Software verification on Target processor or equivalent instruction set simulator 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 3. Model Based Development : V- Cycle Require- SYSTEM. ments Validation HIL. Car Physical Model MvS. (FLP) Integration Test HIL. Implemen- SIL. MvS PIL. tation TVG. Model Module (FXP) Test FLP : Floating point model FXP : Fixed point model MvS : Model verified by Simulation SIL : Software In the Loop Automatic Code PIL : Processor In the Loop HIL : Hardware In the Loop Generation TVG : Test Vector Generation SOFTWARE.

3 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 4. Model Verified by Simulation (MvS). Test Functional Expected cases Requirements Values Modeling . MIL Physical Model Result model(FLP). In the Loop Scaling . Implementation Result Model(FXP). SIL. ACG =. Software In the Loop C-Code Result (s-function in model). =. PIL C-Code Processor (compiled for Result In the Loop C- Target). 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 5. Definition MIL/SIL/PIL. MIL SIL PIL. Model In the Loop Software In the Loop Processor In the Loop Refers to the kind of testing done Refers to the kind of testing done Refers to the kind of testing done to verify the accuracy / to validate the behavior of the to validate the referenced model acceptability of a plant model or auto generated code used in the by generating production code a control system.

4 Controller. using the model reference target. MIL testing means that the The embedded software The code is cross-compiled for model and its environment are is tested within a simulated and executed on a target simulated in the modeling environment model but processor or an equivalent framework without any physical without any hardware. instruction set simulator. hardware components. PIL level of testing can reveal MIL allows testing at early SIL also allows to verify the code faults that are caused by the stages of the Development cycle. coverage. target compiler or by the processor architecture. 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 6. Case study on Engine Temperature function 1 Test suite for calculation of load information for coolant temperature model C_CRLC_LOAD_TCO_MDL.

5 <MAF_KGH>. C_MAF_KGH_MAX_LOAD_TCO m u x_out <LOAD_TCO_MDL>. R. 100% x_in IV = 0. <LV_ES>. <LOAD_TCO_MDL>. 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 7. Model In the Loop (MIL): Floating point model STIMULI. Simulation DISPLAY. project MAF_KGH || <MAF_KGH>. <LOAD_TCO_MDL> || LOAD_TCO_MDL. LV_ES || <LV_ES> ENTE_SIGCVTCOAI. FLP X FLP. Controller/Module Stimuli/Inputs Display/Verify Level 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 8. Model In the Loop (MIL): Fixed point model STIMULI. Simulation DISPLAY. project MAF_KGH || <MAF_KGH>. <LOAD_TCO_MDL> LOAD_TCO_MDL. ||. LV_ES || <LV_ES> ENTE_SIGCVTCOAI. FXP X FXP. Controller/Module Stimuli/Inputs Display/Verify Level 8 July 2014.

6 Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 9. MvS: SDA Simulation Manager Deviations can be detected and can be solved at early stages 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 10. Comparison results MIL - FLP/FXP (Error). Deviations due to wrong scaling 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 11. Comparison results - FLP/FXP (Corrected Case). Deviations are with in the resolution 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 12. Present Situation after MIL. 1 Random test cases are generated to test production code. 2 Execute generated test cases in the project environment.

7 3 More effort is required to prepare test cases to verify production code. 4 Completely different test cases are used to verify model and generated code. 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 13. Wouldn't it be nice to reuse the MIL test cases for test of the Automatically 1. Generated Code ? 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 14. Software in the Loop: SIL. STIMULI Simulation DISPLAY. project MAF_KGH || MAF_KGH. LOAD_TCO_MDL || LOAD_TCO_MDL. LV_ES || LV_ES ENTE_SIGCVTCOAI. SIL SimMode:SIL SIL. Stimuli/Inputs Sfunction block Display/Verify 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 15.

8 Comparison results - MIL/SIL (wrong case). 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 16. Comparison results - MIL/SIL (correct case). No Deviation 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 17. Processor In the Loop: PIL. STIMULI Simulation DISPLAY. project MAF_KGH || MAF_KGH. LOAD_TCO_MDL || LOAD_TCO_MDL. LV_ES || LV_ES ENTE_SIGCVTCOAI. PIL SimMode:SIL PIL. Stimuli/Inputs Sfunction block Display/Verify 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 18. Comparison results - SIL/PIL. 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 19.

9 PIL results for different target processors - Reusability No Deviations 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 20. Conclusion 1 Necessary test effort can be essentially minimized across simulations. 2 Tests suites are portable and reusable. Cost-efficient consistent testing for all phases of the Development : 3. One test suite for all Development phases (MIL, SIL, PIL). 4 Early malfunction detection. 5 Eases the updating of test suites for changed requirements. 6 Shorter Development process resulting in significant time-to-market advantage. 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 21. Thank you for your attention! 8 July 2014. Engine Systems Public Narayanamurthy S, Narendrakumar P, Stefan Schmidt & Ralf Garrelfs Continental AG 22.