Introduction to

PopcornSAR Adaptive AUTOSAR

(R20-11) Tool chain
March. 2022
Index

1. Automotive Software Development Process with Adaptive AUTOSAR (R20-11)

2. Overview of PopcornSAR Adaptive AUTOSAR (R20-11) tool chain
3. Product introduction: PARA
4. Product introduction: PACON IDE
5. Product introduction: Virtual ECU
6. Example of PopcornSAR AP development environment
7. PopcornSAR AP tool roadmap(~2023)

(Appendix 1) Tool demonstrations

(Appendix 2) ISO26262 Certificate of PopcornSAR Adaptive platform Authoring tool
1. Automotive Software Development Process with Adaptive AUTOSAR (R20-11)
1. PopcornSAR provides automotive software development tool chain for Adaptive AUTOSAR Applications(AA).
• Supporting the whole development process of Adaptive Applications/ECUs for OEM/Tier1.
2. Able to automate the whole development process based on CI/CD after the initial manual development.
3. Yocto SDK is used for Adaptive application(AA) development after the SOP.

Initial manual development

CI/CD pipeline
E/E Architecture design SW Implementation Integration System test ECU test
C++code/Manif Platform-level AAs Function Function
Requirements AUTOSAR Yocto SDK
(EM/SM/CM/S2S/ verification verification
modeling est Generation
PHM/DM/UCM) (Virtual ECU) (Physical ECU)
Analyzation Executable
Result [Results : ARXML] [Results : Manifest(json)
[Results]
/ Codes(Gen.)]
AA verification
- Application Design Report
- Machine Manifest - Application Design [Results]
- Execution Manifest - Machine Manifest
- FCs library (PARA)
- Service Instance - Execution Manifest
Manifest etc. - Service Instance
Manifest etc.
- Manifest (*.json) POSIX OS
- ARXML (configuration
(Execution)
management)
Application-level
AAs
Jenkins CI/CD
Tool : AutoSAR.io AA - Yocto Recipe
[Results]
verification
Report
BSP Image
(Docker Container)
$ Docker Run => Execute Execution Management(EM) PopcornSAR’s provision
Tool : PACON IDE
(Including PARA)

3
1. Automotive Software Development Process with Adaptive AUTOSAR (R20-11)

ARXML C++ code/Manifest

User Code Compile Virtual ECU
Design Generation

Requirements
Manifest Files
Hand-coding
Generated C++ Codes LIB FC library files
Codes After Generation

Executable Files

• After testing on virtual ECU, developers need to conduct system test on

the physical ECU.

Develop SOP Shipment Disposal

AP version (R20-11) Upgrade AP version through OTA (R21-11)

• Vehicle functions and cyber security can be updated through OTA after SOP.

4
2. Overview of PopcornSAR Adaptive AUTOSAR (R20-11) tool chain

• A development platform with tools installed in PC

① SDK of Adaptive Platform(AP) Functional Cluster(FC) or Vehicle APIs (No dependency on middleware vendor).
② Providing virtual ECU (simulation & test), supporting programing languages like C＋＋.
③ Coding Rule check, API autocomplete, etc.

Dev. server PARA

AutoSAR.io Vscode PACON IDE (User N)
vscode-server Vehicle SDK
Compiler other tools

Virtual ECU Docker Container N

System Test Docker Container N

Not provided by Ubuntu(HOST OS)

PopcornSAR
ARXML modeling tool
Adaptive Platform Functional
(Supports AP & CP)
Cluster(FC) & Platform Level
Adaptive Applications

5
3. Product Introduction : PARA - FCs list (R20-11) To-be (before 2023)

AUTOSAR Runtime for Adaptive Applications (ARA)

EXE EXE EXE EXE Py
ara::com ara::tsync Libpara Python
ara::per ara::diag ara::sm
communication
persistency
time
diagnostics state management
Library of safety SOME/IP
management synchronization mechanism AI Python for SOME/IP

EXE EXE EXE EXE EXE Py

ara::core ara::exec ara::phm ara::log ara::nm S2S CAN-DO

execution platform health network Signal-to-service
core types logging & tracing CAN virtual ECU
management management management management

EXE EXE EXE

ara::iam ara::crypto ara::ucm AA for NON-AA
identity access update & configuration
cryptography EM/SM for non-autosar
management management

POSIX PSE51/C++ STL

operating system interface
(Virtual) Machine / Container / Hardware

Available Available C++ Source Code Manifest Specific Platform Level PopcornSAR
EXE
now before 2023 Generator Generator Generator Application Specific

6
3. Product introduction: PARA - Features

1. PARA enables Adaptive Application tests without Execution Management(EM).

• According to AUTOSAR standard, EM is essential. However, PARA users can
develop/test adaptive applications without EM, increasing development efficiency.
2. If no problem is found in ARXML Validation, no more parameter is needed for
PopcornSAR’s generators, and the generation will be done automatically.
• AutoSAR.io includes generators of FC, so it automatically analyzes Adaptive
Applications and Machine, and generates related source codes and Manifest files.
3. PopcornSAR can migrate results(C++) from Matlab/Simulink 2021a(R19－11) or AP M
odel Based Design(MBD), Legacy MBD into Adaptive applications, and related
guidelines are also available.
• It’s necessary to use Matlab/Simulink to develop Adaptive Applications for control.
4. PARA supports Python’s SOME/IP for AI development.
7
3. Product introduction: PARA - Features

<Generation finishes in average of 1 minute by PARA in AutoSAR.io>

8
3. Product introduction: PARA – Safety Mechanism for ISO26262

• PopcornSAR provides special safety mechanism(libpara) which is developed by

PopcornSAR own. It can detect faults in execution level when AA is running, such as:
1) Human errors in using ARA API when implementing Adaptive Applications,
2) Human errors in using POSIX OS.

<ex1 Human error of ara API : SOME/IP Error fault detection when AA runs: someip OfferService API>
2022/02/09 03:39:18.645278 102307244 001 ECU1 CM-- DFLT log error V 1 [InitByManifest:: ManifestParser::InitByManifest::
HasParserError fail <0> (/home/popcornsar/para-r2011-main/para-api/com/internal/database/database.cpp #63)] 5terminate
called after throwing an instance of 'std::runtime_error' 6 what(): InitByManifest:: ManifestParser::InitByManifest::
HasParserError fail <0> (/home/popcornsar/para-r2011-main/para-api/com/internal/database/database.cpp #63

<ex2 Human error of POSIX OS : dummy files created when using vi editor of POSIX OS in target board>
9
 3. Product introduction: PARA – CAN-DO(based on Docker container)

• CAN-DO can generate CAN virtual ECU which is based on docker container with
ARXML/DBC/EXCEL.
• Able to fast generate multiple CAN virtual ECUs for testing ECUs.
• Able to manage CAN virtual ECU’s configurations through Dockerhub, and able to
automate the virtual ECU generation process through Jenkins.
• No limitation of AUTOSAR versions.

Input Generation Output

<ARXML / DBC / EXCEL> <CAN Virtual ECU based on Docker container>

<.CSV or .json>

Configuration
Auto-distribution and generation through Jenkins management

10
3. Product introduction: PARA – Example use case of CAN-DO
ADAS ECU for Variant 1 ADAS ECU for Variant 2

AA1 AA1

AA2 Adaptive AA2 Adaptive

Vehicle variant 1 AA3
Application
Variant 1 Vehicle variant 2 AA3
Application
Variant 2

…
ECU P-1 ECU C-1 ECU P-I ECU C-I
Powertrain Chassis Powertrain Chassis
virtual ECU P-2
… virtual … virtual ECU P-II
… virtual …
ECU C-2 ECU C-II
CAN ECU CAN ECU CAN ECU CAN ECU
Variant 1 ECU P-3 Variant 1 ECU C-3
Variant 2 ECU P-III Variant 2 ECU C-III

CAN

• Virtual CAN ECUs can be generated by CAN-DO automatically.

• Configurations can be saved & managed by Dockerhub/Docker registry.
3. Product introduction: PARA – AA for Non-AUTOSAR Applications execution

AA1 AA2 AA3 Management Container A Container B

& Execution
Non- Non-
SM Legacy network EM/SM for AUTOSAR AUTOSAR
Non- AA1 AA2
EM CAN / LIN AUTOSAR (Python) (Python)

Host OS AA Docker Container

(Adaptive AUTOSAR) (Non-AUTOSAR)

• EM might not recognize non-AUTOSAR Apps (docker, legacy network app, Python etc).
• PopcornSAR provides special AA as a role of EM/SM to execute Non-AUTOSAR Apps.
4. Product introduction: PACON IDE – Features

1. PACON IDE(=Docker Container) can be customized according to customer’s needs.

• Separate IDEs for different variants.
• Customer’s IDEs can be redistributed & managed through Jenkins CI/CD.
2. Has no dependency on Adaptive Platform Vendor.
3. Code developer can use Vscode to connect with PACON IDE.
• Docker Container has essential extensions for using PARA & Vscode.
• External network is not needed.
4. Wireshark is provided even if it’s not installed in developer’s PC.
5. Can be distributed through Jenkins CI/CD.
6. Provides additional features for efficient development.
• Auto complete function for ARA API & open source API.
• Real-time code rule check.
• Debug.
• Various programming languages (Python etc.,).
13
4. Product introduction: PACON IDE – Two variations of PACON IDE

① Develop within developer’s PC ② With connection to development server

(Without external network) (Increase convenience & efficiency based on
remote co-working environment)
Developer’s PC
Developer’s PC Dev. Server
PACON IDE PACON IDE (User N)
VSCode-server Vehicle SDK VSCode-server Vehicle SDK
Compiler other tools Compiler other tools
VSCode VSCode
Virtual ECU Docker Container Virtual ECU Docker Container N

System Test Docker Container System Test Docker Container N

Windows/Ubuntu/Mac Ubuntu(HOST OS)

Windows/Ubuntu/Mac

docker container：Distributed through Dockerhub

14
4. Product introduction: PACON IDE – Customization according to requirements

Variant CPU OS
Variant 1 NVIDIA Linux (5.0) Variant1:1.0.0
NVIDIA Linux (5.15) Variant1:1.1.0
Variant 2 Intel QNX Variant2_intel:1.0.0
NXP QNX Variant2_nxp:1.0.0
Variant 3 R-car Android Variant3:1.0.0

• Separate IDEs for different variants.

• Customer IDE can be redistributed & managed through Jenkins CI/CD.
• IDE must be redistributed according to updates of OS. Auto-distribution
through Jenkins

15
4. Product introduction: PACON IDE - IDE and Virtual ECU creation process

Development environment
Information of for target ECU (=PACON IDE)
customer’s target ECU
vscode-server Vehicle SDK
CPU
(NXP, RENESAS, Compiler other tools
NVIDIA etc.)
Virtual ECU corresponding
POSIX OS to physical ECU
(Linux, QNX etc.) Virtual ECU Docker Container N Configuration
management
Container for system test
Compiler
Functional clusters System Test Docker Container N

16
5. Product introduction: Virtual ECU – Features (1)

1. PopcornSAR’s virtual ECU can be customized according to customer’s requirements.

• Separate virtual ECUs for different variants.
• Customer’s virtual ECUs can be redistributed & managed through Jenkins CI/CD.
2. Virtual ECU has no dependency on Adaptive Platform Vendor.
3. PopcornSAR’s Virtual ECU has lighter footprint, inducing efficient development with less
compatibility issues than QEMU.
• PopcornSAR provides virtual ECU(docker container) which corresponds to customer’s physical ECU
requirements(POSIX OS etc.).
• After being tested on Docker Container, adaptive applications can be copied directly on target ECU.

17
5. Product introduction: Virtual ECU – Features (2)

4. Developers can generate virtual ECUs through PACON IDE.

5. Developers can develop multiple ECUs in docker container, manage their configurations through
Dockerhub or Docker registry, and share their results with other colleagues.
6. Multiple virtual ECUs can be generated by Jenkins CI/CD for adaptive application tests.
7. PACON IDE provides additional docker containers for virtual ECUs to carry out other system tests.
• Usually, it’s difficult to add system testing software in virtual ECU. (Ex : Installing tshark in virtual ECU ARM)
• PopcornSAR can generate additional docker containers for system testing software.

18
5. Product introduction: Virtual ECU – Customization according to requirements

• IDE and virtual ECU for the same variant are managed separately.

Variant CPU OS
Variant 1 NVIDIA Linux (5.0) vECU_Variant1:1.0.0
NVIDIA Linux (5.15) vECU_Variant1:1.1.0
Variant 2 Intel QNX vECU_Variant2_intel:1.0.0
NXP QNX vECU_Variant2_nxp:1.0.0
Variant 3 R-car Android vECU_Variant3:1.0.0

• Separate virtual ECUs for different variants.

• Customer’s virtual ECU can be redistributed & managed through
Jenkins CI/CD.
• Virtual ECU is compatible with PopcornSAR’s/other ECUs. Auto-distribution
through Jenkins
19
5. Product introduction: Virtual ECU – Identical environment to target ECU

Dev. server Target ECU

Virtual ECU Docker Container
Adaptive Application 1
Adaptive Application 1
Adaptive Application 2
Adaptive Application 2
Adaptive Application 3
Adaptive Application 3
NXP S32G ARM POSIX OS
Same BSP Image
NXP S32G ARM POSIX OS
Ubuntu(X86 HOST OS)

In case of QNX, virtual ECU will run on QEMU instead of docker container.

• Once testing is done in a development server, an adaptive application can be ported to target ECU
without recompiling.
-> Maximizes efficiency in development.

20
5. Product introduction: Virtual ECU - QEMU vs Docker Container

QEMU (AS-IS) DockerContainer(To-be)

Real-time simulation for
X
adaptive application in target O
(Consuming ROM/RAM)
board
Virtual ECU
X O
configuration management
Sharing with others X O

Distribution through network X O

Adding adaptive application

X O
in virtual ECU

Can connect to other docker

Others -
containers for system tests

21
5. Product introduction: Virtual ECU - Examples

• Developer can test network by generating multiple virtual ECUs.

ARM NXP S32G ECU1

ARM NXP S32G ECU2

ARM NXP S32G ECU3

<Virtual ECU in PACON IDE> <Terminal in PACON IDE>

22
5. Product introduction: Virtual ECU - Examples

• Can use tshark in Docker Container for system test to monitor virtual ECU(ARM, etc.).
• Can use Wireshark which is installed in PACON IDE.

23
6. Example of PopcornSAR AP development environment
• PopcornSAR AP tool chain doesn’t include QEMU nor Yocto.
• Since ARA API changes frequently for each AP version, PopcornSAR developed its own testing SW.
Dev. server
<Management through
AP Docker Container 1 Sharing Simultaneously
jira & confluence>
vscode-server Vehicle SDK
Compiler Virtual ECU Docker Container N
other tools

＜ Dev. Leader＞ System Test Docker Container N

Tshark, Python, etc.

AP Docker Container 2
vscode-server Vehicle SDK
Compiler other tools Volume

＜FC Developer＞
AA source code
& build files
for System Test
AP Docker Container 3
vscode-server Vehicle SDK
Compiler other tools

＜Test Assistant＞

24
7. PopcornSAR Adaptive tool development roadmap (~2023)

▶R20-11 ara::ucm
▶R20-11 ara::phm ▶Signal Management
▶R20-11 ara::diag (like S2S) ▶PARA QNX ▶R20-11 ara::crypto
▶PACON IDE & vECU ▶python SOME/IP ▶R20-11 ara::nm ▶R20-11 ara::iam
(QEMU) for QNX library ▶R20-11 ara::com::e2e ▶R20-11 ara::tsync
1Q 2Q 3Q 4Q
2022

▶Available for R21-11 & R22-11 ▶Start developing R22-11 ▶Provided according to PARA

2023

AutoSAR.io PARA PACON IDE

25
(Appendix 1) Tool demonstrations

1. R20-11 Autosar.io ARXML modeling :

- Link : https://youtu.be/3FYzR0bQ44s

2. R20-11 generation & build for communication between

2 Adaptive Applications :
- Link : https://youtu.be/HYqNEMrYYAw

3. R20-11 ARA::DIAG demo:

- Link : https://youtu.be/jvySoUdoAJs
- Link : https://youtu.be/tm_Cr80d52w

26
(Appendix 2) ISO26262 Certificate of PopcornSAR Adaptive platform Authoring tool

27