Innovation of automotive software development · remote touch controller, smart key, advanced...

This information is the exclusive property of DENSO CORPORATION. Without their consent, it may not be reproduced or given to third parties.

Masahiro Goto Director ePF R&D Division DENSO CORPORATION

Innovation of automotive software development


Corporate Profile

December 16, 1949 Established

Capital 187.4 billion yen (US$1.8 billion)

Notes: U.S.dollar amounts have been translated, for convenience only, at the rate of 102.92 yen = US$1, the approximate exchange rate prevailing on March 31, 2014. Billion is used in the American sense of one thousand million.

Operating Income

377.7 billion yen (US$ 3.7 billion) 223.3 billion yen (US$ 2.2 billion)

Consolidated basis Non-consolidated basis

(Japan 13, North America 4, Europe 3, Asia/Oceania 11, South America/Others 2)

Affiliates under the equity method 33

(Japan 62, North America 26, Europe 35, Asia/Oceania 56, South America/Others 6)

Consolidated subsidiaries 185

Employees

139,842 38,581


Net sales

4,095.9 billion yen (US$39.8 billion) 2,490.8 billion yen (US$24.2 billion)


/ as of March 31, 2014

*There are two non-consolidated subsidiaries.

Profile 1 / 27


Comfort & Convenience

Hybrid and electric vehicle components,

gasoline engine management system,

diesel engine management system,

starter, alternator, radiator, etc.

Safety Sensing technologies for driving assist systems,

actuator & computer for antilock brake system (ABS) /

electronic stability control (ESC),

adaptive front-lighting system (AFS),

airbag sensors & electronic control units,

periphery monitoring system, instrument cluster,

rain sensor for automatic windshield wiper, etc.

Car air-conditioning system,

air conditioner for buses, air purifier,

Car navigation system,

electronic toll collection system (ETC),

remote security system,

remote touch controller, smart key,

advanced vehicle operation system (AVOS), etc.

Environment

Automotive Fields 2 / 27


Challenges surrounding Mobility Society

Energy collaboration with homes and local communities PHV,EV to Smart grid, Micro grid

Semi-automatic drive by coordination with infrastructure （Vehicle-vehicle, Road-vehicle） Utilization of IT・Cloud by broadband connection Change of the usage style of the car

Car sharing, Multi modal

Future Society

Realize Zero-traffic accident society Pre-crash safety ～ Active safety～

Drive assist Make existing products low cost for

wider spread and develop further advanced safety device

Traffic Safety

Contribution to developing an environment friendly to the earth

CO2 reduction, fuel efficiency improvement and diversification of energy sources to prevent global warming

More electric（Hybrid, PHV, EV） Change from Engine to Motor

Global Environment

PHV: Plug-in Hybrid EV: Electric Vehicle

3 / 27


140

130

95

70

0

30

60

90

120

150

0%

20%

40%

60%

80%

100%

2010 2015 2020 2025

Stay in

CO2 =450ppm

Global environment : Global warming and counter measures

Share

New Policies Scenario *1

Forecast

2015 2020 2025 2030 2035 2010 2010 2015 2020 2025

CO2 free / Low CO2 emission vehicle will expand

Conventional ISS CO2 of new car

HV EV

[Gt]

[g/k

m]

[Source] World Energy Outlook 2012, IEA.

[Source] Share forecast : DENSO

CO2 of new car: Ricaldo

*1 assuming most energy consuming countries implement new

policies of energy effciency

Forecast of CO2 emission Powertrain Mix (450ppm scenario)

4 / 27


0

200

400

600

800

1000

1200

0

2

4

6

8

10

12

14

16

18

1965 1975 1985 1995 2005

1225.9 1170.7 （’98） World

5.3（’98） Brazil

84.6 32.2 SE Asia

74.7 India

73.7 China (PRC)

30.4 55.5 EU 25

35.5 42.1 U.S

4.9 9.9 Japan

trend 2010 1996

276.0

231.0

43.9

Fata

litie

s (t

housa

nd)

• stronger crackdown,

safety education

• Infrastructure improvements

• seatbelt obligated

• front-crash test obligated

• Airbag, ABS 100% installed

[Source]

Based on various statistical data in 2009

(Exhibited by: Global status report on road safety (WHO), European Road Statistics 2009 (IRF), Transportation conditions report of major countries(MLIT), NPA, NHTSA, The Ministry of Public Security of the People's Republic of China.)

Global Health Observatory Data Repository(WHO) 2014 WHITE PAPER ON TRAFFIC SAFETY IN JAPAN (2004, 2014)

*1 25 countries (excepted Bulgaria, Romania , Croatia from EU-28)

*2 Philippines , Malaysia , Singapore , Indonesia , Thailand

Safety - traffic accident fatalities in global and Japan

*1

[Source]

WHITE PAPER ON TRAFFIC SAFETY IN JAPAN (2004, 2014) AIRIA

*2

(thousand) Traffic fatalities in global

Emerging country is similar to 1960s Japan ⇒ Demanding Airbag, Seatbelt, ABS/ESC

Airbag, ABS/ESC already spread in JP,US,EU ⇒ Demanding ADAS (Advanced Driver Assistant)

Traffic fatalities in Japan

Num

be

r o

f a

ccid

ent

(tho

usa

nd)

Num

be

r o

f ve

hic

le (m

illio

n)

5 / 27


Anyone, Anytime, Anywhere : enjoy various services through network without any stress.

Future Society ：Mobility society image

Ubiquitous Society

Smart grid Widespread use of IT services

Social control Automated driving

POI

map

Route calculation

・・・

cloud

permanent connection

Broadband connection

New mobility society in cooperation with social infrastructure

BATT

BATT

Monitoringsmart meter

BATT

BATT

Monitoringsmart meter

Smart phone

cellphone

network WiFi

V2I

V2V DSRC

generation

generation transmission

PLC or radio

Multi-modal mobility

Transit fee

Car sharing

Park & Ride

Distance control

Crash avoidance

Pedestrian detection

Charging spot

Expected

travel distance

Operation &

management

center

management

V2V: Vehicle to Vehicle communication V2I: Vehicle to Infrastructure comm. DSRC: Dedicated Short-Range Comm.

PLC: Pow er Line Communication

Controll Center

6 / 27


Why does Automotive systems become complex?

Driving Context

Physical Phenomena

Human Operation

System requirements are developed through

iteration of prototyping and evaluation

3000枚

8900枚

96年 98年 05年

仕様書枚数

650枚

仕様書枚数（制御系システム）仕様書枚数（制御系システム）仕様書枚数（制御系システム）

Automotive Electronic System Development

Control Specification

• Spec. becoming too detailed and complex.

• Intent getting harder to understand

7 / 27


Automotive embedded software development

Incre

asin

g va

riatio

ns d

ue to

car lin

e

Increasing number of versions due to functional improvements

A large number of similar but different software is developed

08MY-大衆車

Functionality A-3.2

08MY-大衆車

Functionality A-3.1

03MY-中級車

Functionality A-2.2

06MY-中級車

機能A-2.1Functionality B-2.2

09MY-中級車

機能A-2.2Functionality B-2.1.2

03MY-中級車

Functionality A-2.1

06MY-中級車

機能A-2.1Functionality B-2.1

09MY-中級車

機能A-2.2Functionality B-2.1.1

02MY-高級車 05MY-高級車 08MY-高級車

Functionality A-1Functionality B-1.1

機能A-1.1Functionality -1.1

機能A-1.2

Functionality C-1.1

01MY-flagship 04MY-flagship 07MY-flagship

Functionality A-0 Functionality A-0.1 Functionality A-0.1.1Functionality B-0 Functionality B-0.1

Functionality C-0

08MY-Ordinary car

Functionality A-3

02MY- High-class car Vehicle 08MY- High-class car

Functionality A-1Functionality B-1

Functionality A-1.1Functionality B-1.1Functionality A-1.2

Functionality C-1

03MY-Mid-class car

Functionality A-2

06MY-Mid-class car

Functionality A-2.1Functionality B-2

09MY-Mid-class car

Functionality A-2.2Functionality B-2.1

8 / 27


ADAS (Advanced Driving Assistance System)

Reduce Traffic Accidents

Damage reduction

in case of car crash

Pre-Crash

・ Low-speed → High-speed ・ Pedestrian ・ Lane-kpeeping → Off-lane ・ Reduce driver’s load ・ Reduce human error

Infrastructure / Crowd

・ Intersection accident avoidance ・ Automatic driving

Driver

Vehicle

Technolo

gy E

xpansio

n

Influence for Society

Society

source: ITS World Congress @Oct. 2013

Pedestrian PCS (Toyota)

Prevention mistake of gas pedal (Nissan)

City brake active system (Honda)

“We are developing driving support systems to reduce the danger

in an emergency (safety when needed) and to take away drivers’

worries (peace of mind always).”

http://www.globaldenso.com/en/investors/library/annual_report/2013/index.html

9 / 27


Camera

MMW RADAR

LIDAR

Ultra sonic

Wireless com.

Driver Monitor

Application

Prevent misacceleration

Adaptive Cruse Control

Collision Mitigation

Auto Parking

Collision Warning

Auto Lighting Control

Lane Departure Warning

・・・

Brake

Engine

Light

Driver

Safe

ty M

anager

HM

I Manager

Ve

hic

le M

otio

n C

on

trol

Instrument

Cluster

Engine

Center Display

ESC

Light Control

EPS

ECU

Senso

r Manager

Steeling

ADAS as a complex system

Integrating sensors and actuators to provide various services

Service

10 / 27


Context of ADAS

Various road environment over the world

depending on traffic situations, customs, and cultures

11 / 27


Target Demonstration

Engine Unintended acceleration

Brake Unintended deceleration

Steering Unintended steering

Indicator Indicating 100mph even when stationary

Horn Sudden beeping

Seat belt Sudden tightening of seatbelts

Gas gauge Inaccurate indication of gas

source: DEF CON 21 @Aug. 2013

Cybersecurity

Evolving Services • Connected vehicles

• ADAS

Negative aspect : Cybersecurity • Demonstration of carjack (DEFCON)

12 / 27


In-Vehicle Systems

Target System

Sa

fety

Asset

Pro

pe

rty / P

riva

cy

Connected Systems

Powertrain

Chassis

Body

ADAS

Quality

Multimedia Reprograming Security

Information Security Car Security

Functional Safety

Influence for Society

Evolving Automotive Electronics getting more critical

13 / 27


An Issue on the Automotive System Development

• Growing complexity of the systems

• Diversifying the system contexts

Major following observations

indicate the importance of

Getting higher confidence of

automotive electronic systems

Confidence is the source of both:

• responsibility to provide high quality of the products, and

• accountability to promise peace of mind to the society

14 / 27


Calibration & Validation

Calibration tool

Calibration DB

Specification development & Verification

Control model description tool

Control model verification tool

Existing Solution

Automated test bench

制御モデル

C program

Automatic code generation tool

Control model Code generation & test

実装

Design

Verification

Implementation

Automatic calibrator

Model-based development Simulation technologies promote automation of development

15 / 27


Why “Testing” is important

It is very important to verify “it works on-specification”

“It works on-specification” does not only

mean the system satisfy the performance

written in the document,

but also includes the system will not do any

behavior other than written in the document.

Thus, it is much important to consider various conditions and

environment (even when it is un clear),

and check if the system actually works using various test case.

16 / 27


Concept Validation Requirement Elicitation Standard Asset Development Product Development

Architecture Design

Component

Design

Integration

Testing (SW)

Unit Testing

ECU Integration

Testing

Acceptance

Testing

System

Testing

System Requirement

Definition

HW Design HW Design

Control Design needs

System Plan

Business needs

Plant model

Simulation

Prototype Functional model

Technical model

Behavior model Verif ied

Components

System

Test spec

ECU test

Spec.

Functional

Non-

Functional

Requirement

Requirements

are defined

along with

development

responsibility

Control model

Unit

Implementation

Typical example

“Testing” during Development Lifecycle

• unit testing in software creation

• Integration testing in system development

• System testing on early prototype phase

17 / 27


Testing in software unit level

Back-to-back test between Model and Software • Compare behavior based on the structural coverage

TestVector

Model Simulation

Instruction Set level Simulation

Model Sim.resultOut1

Code Sim.resultOut1

Comparisonvoid untitled_obs_step(void)

{

int16_T rtb_Switch;

if (In2) {

rtb_Switch = In1;

} else {

rtb_Switch = In3;

}

if (In4 < rtb_Switch) {

rtb_Switch = In4;

}

Out1 = rtb_Switch;

}

18 / 27


Vehicle Level

Function Level

Control Level

Implementation Level

Function

Design

Req.

Analysis

Functional

structure

Control

Design

discretization Evaluation

Implementation

Integration spec

Coding

Source code

Unit Testing

Integration Testing

Vehicle system

Calibration

Calibrated module

Calibration

constants

Validation

Functional

Test Scenario

Functional

Test DesignState transition

condition

Requirements

Control logic

Plant model

Control

spec

Impl model

Tested source code

Integrated module

Evaluation

Calibration

procedure

Testing in software unit level

High automation is available

Gap is close

• Test Execution can be automated

• Test Vector can be automatically generated

• Less possibility to find bugs

• High confidence against coverage

But effect is limited

Measures to prevent creating bugs like

restricted coding rules and auto code

generation are more effective.

Peer review is more efficient to find

coding bugs.

Bugs of compilers are only found by

testing.

19 / 27


Testing for integration* and functionality

Requirement based testing • Compare Implementation and Requirement

* SW Integration, HW/SW Integration, System Integration

Stop

Back

Wait Start_avoid

Finish_stop

Finish_avoid

Finish_avoid

>0

Conditions to cover Requirement

Map to Implementation

Implementation Model

Abstract model (Observer) Test cases

Expected Behavior

20 / 27


Testing for integration and functionality

Automation is limited • Test Execution can be automated

• Automated Test Generation does not scale

Decompose

Requirement ModelRequirement

Modeling

System

Design

Component

Design

Testing (simulation)

Refinement

Requirement Model

Impl. Model

Test Generation

Test Data

Testing (Target)

• More work than test case design

Modeling is intensive

Simulink models used for product development

contain more than ten thousands of blocks and

take several thousands of simulation steps

21 / 27


System testing on early prototype

Coolant circuit

HVAC

Batt

Electrical conpopnent

Electrical circuit

Transmission

Engine

Coolant circuit

HVAC

Batt

Electrical conpopnent

Electrical circuit

Transmission

Engine

0

100

200

0

2000

4000

0

2

4

-50

0

50

100

-20

0

20

Vehicle speed

（km/h）

Engine

revolution

（rpm）

Fuel

injection

（g/sec）

Coolant

temperature

（℃）

Cabin

temperature

（℃）0 200 400 600 800 1000 1200

Time （sec）

Experiment

Simulation

Prototype Evaluation

Plant Modeling

Control Design

Collect data

Revise model to fit required accuracy

Revise control

using simulation

How testing technology assist early phase?

22 / 27


Test Selection (Completeness)

Test O

racle (Co

rrectness)

SufficientInsufficient

No

t Accu

rate

Accu

rate

MBT Asset management

Regression, …

Combinational methodAll-pair, HAYST method, …

Ideal Proof“Fundamental” Formal Method

Monkey

Test as “Evidence” of verification

Test to find Bugs

Requirement-based

Verification

No strategy

Testing Strategy

Control trade-off in accordance with development phase and purpose

23 / 27


Findings So Far

• Simulation technologies are widely used

• Test generation technology is available

but effect is limited

• System is getting more complex, which causes:

Scalability is crucial for the tools

Modeling is required as engineers’ work

24 / 27


Learning curve for C Coding

Learning curve for Modeling

Performance

How Technologies impacts Organization

Half Year (approx.)

Many Years (not predictable)

Time

Acceptable level

for job

Technologies require engineers’ skills

• Introducing Automation Understanding both Technologies and Process

• System Design Modeling domain specific requirements

25 / 27


Human resource capability model

Exce

rsiz

e

Outcome

Knowledge ・Technology

Basic skills

Values・Mind

Personality・Fundamental ability

Late

nt

Ove

rt

Conceptualization skill

Personal self-consciousness

Su

pp

ort

Phase

Are

as

Softw

are

Syste

m

Planning・Research Development Mass production

Process improvement・Quality management・ Human development・ Intellectual property management

Generate profits by

producing products

Develop

Standard

assets

Sow

seeds

Development know-how

0

200

400

600

800 Understanding

Analysis

Conceptualization Relationship

Leadership

Domestic Overseas

Management of diverse human resources

Clarify “What an employee wants to

do”, “What an employee can do”,

and globally implement “right person in

a right place”

26 / 27


Conclusion

• Automotive technology is evolving rapidly and brings various influence

on the society.

• It is a key to get higher confidence of automotive electronic systems.

I would like to emphasize the importance of academia-

industry collaborations

• to see and to understand what is happening in gemba, or

the actual place, and

• to find out the essence of the issues and to improve

appropriate technologies for the solution !

27 / 27

Many excellent idea and technologies are proposed from academia, but it

looks like only few of them have been applied as industrial solutions

because of various constraints and issues.

Innovation of automotive software development · remote touch controller, smart key, advanced...

Documents

Transcript of Innovation of automotive software development · remote touch controller, smart key, advanced...