From: AAAI Technical Report SS-99-04. Compilation copyright © 1999, AAAI (www.aaai.org). All rights reserved.
Integrated Diagnosis for Future Mobile Systems
Technology Overview and Concepts
Dr.-Ing. B. Baker,Dipl.-Ing. Th. Forchert
DaimlerChrysler AG
Researchand Technology,FI’2/EK
D - 70546Stuttgart, Germany
bernard.baeker@daimlerchrysler.com
thomas.forchert@daimlerchrysler.com
Abstract
This paper describes newhardwareand software diagnosis
conceptsfor OnBoard
use in mobilevehicles. Followinga
review of existing diagnostic methodologies,we discuss
approachesto the developmentof next generation modelbased diagnosis tool environmentswhich1) automatethe
productionof diagnosis modelsfromdesign models,and 2)
can improvethe diagnostic modelingprocess flow from
design through to in-field service and maintenance.
Examples of a model-based reasoning approach to
mechanical(powertrain)and electronic (electrical vehicle
network)component
diagnostics will be shown.
workshops with PC-based diagnosis and centralized
database access;
-
Design and implementation of new AI diagnosis
algorithms (see figure 4-7), including model-based
diagnosis strategies,
enhanced process chain
integration - from the creation of diagnosis data to
the development of diagnosis systems, the use of
overriding software concepts, and the definition of
interfaces for decentralized diagnosis modeling;
Introduction
Global car and truck manufacturers are substituting
increasing numbers of microprocessor-controlled
mechatronic systems and components for purely
mechanical devices in an effort to achieve improved
control and performance at reduced cost [B~ik962]. This
trend requires the adoption of new integrated-vehicle and
component-specific design and diagnostic modeling
methodsthat scale well with increases in the structural,
functional and configuration complexity of mechatronic
command/control systems (e.g. powertrains, suspension
systems, and "X-by-wire" systems), and of new in-vehicle
high speed data/control networks. At the organization
level, product design, manufacturing and diagnostics
service teams need new approaches to collaboration,
communications and learning which support common
knowledge representations
and concurrent planning,
engineering and test across the full vehicle life cycle (see
figure 1: introduction slide).
The shift to an OnBoardnetwork infrastructure (see
figure 3), comprising field bus lacing [Bak973] of
control systems and internal
networks, new
diagnosis-suitable Smart Power sensor technology
elements, and the use of powerful micro processors;
Optimization of the diagnostic modeling process
chain (see figure 8-9), including automation of the
diagnosis data generation process, the diagnosis
algorithm and the needed structure data to diagnose
components and functions, development of modified
software tools environments by vehicle suppliers and
manufacturers, improved data handling, and the use
and cataloging of mechatronic data models;
Requirementsto integrate new system functions with
product value creation (see figure 11), such as new
safety- and driver assistance-functions, and remote
diagnosis- and maintenance-functions[B~ik981].
These system limitations
strongly motivate the
¯ development of integrated, system-wide approaches to
vehicle OnBoardand OffBoard diagnosis.
Motivation
Goals and Benefits
Efficient diagnosis of mobile vehicle systems faces
substantial constraints in the future:
The use of a special OffBoardservice infrastructure
(see figure 2) whichintegrates links to global service
and customer assistance networks and centers,
requirementsfor availability, and the installation of
The fundamental goals and advantages of a system wide
vehicle diagnosis are:
-
Control of networkedelectric/electronic
vehicle.
units in the
Enhanceddiagnosability of system components.
Improvedpassenger safety and vehicle availability
(up-time).
Mathematics and Physics Background
The core integrated OnBoard diagnosis systems (see
figure 4) utilize complex diagnosis algorithms based on
new artificial
intelligence
(AI) processes under
development. These algorithms require knowledge of the
underlying technical system structure (e.g.dependencies
between the internal system values and the system
functions). In principal, qualitative and quantitative
diagnosis strategies will be employed(see figure 5-7)
these procedures portray different attributes and views of
the target system, makingtheir application conditional on
the unit which is monitored.
Automated Data Generation
for Component-Specific Diagnosis Model
Fragments
The development of new diagnosis procedures also has a
large impact on the software tool environment employed
during system design through to production (see figure 89). Complexelectronic structures are today diagnosable
only manually. Consequently powerful strategies are
necessary to generate automatically parts of the needed
diagnosis routines from the available design data.
System Diagnosis - Outlook
An demonstration
implementation
of a ,,System
Diagnosis" illustrates the possibility of future diagnosis
solutions [For981] for mobile vehicles (see figure 10).
With the help of product-integrated diagnostic processes,
internal electric/electronic infrastructure and functioncomponent hierarchy information can be employed as a
basis for the generation of modelsfor localize faults and
fault type, failure causes, and functions which interfere
with user operation. A subsequent degree of severity
further enables the specification
of the procedure
instructions.
As the number of mechatronic system components in the
vehicles of the future continually increases, newdiagnosis
concepts
and technologies
of maintenance and
surveillance of these units will be important as well (see
figure 11-12). These mechatronic units include passenger
compartment and other interior electric/electronic
components, Powertrain electromechanicai components,
"X-by-wire" technology for brake-, control- and damping
systems, and new forms of telecommunication systems.
Conclusions
The application
of Onboard model-based diagnosis
systems in future mobile systems demonstrates numerous
advantages over conventional diagnostic approaches. In
addition to optimizing the process chain with an
integrated tool environment, improvements can also be
achieved in the specification of fault information and
diagnosis deep models [B~ik982]. Together with new
forms of diagnostic infrastructure in customer service and
Offboard vehicle services, customer acceptance and the
vehicle availability can be improved.
References
Book with Multiple Authors
[B~ik973] B~iker, B.; Varchmin,J.-U.: "Simulation intelligenter Kfz-Bordnetze mit CAN-Feldbusanbindung und
Leistungshalbleitern mit integrierter Sensorfunktion", in
4 u" SICAN Autumn Conference,
Proceedings
ISBN 3-9805472-2-X,
SICAN GmbH Hannover,
Hannover - Germany 1997.
Proceedings Paper Published by a Society
[B~tk962] B~iker, B.; Rech B.; Linder W.; Varchmin, J.U.: "Electrical Controlled Generator Drive with an ERFluid Clutch", 5" International
Conference on New
Actuators,
Proceedings Actuator 1996, Bremen Germany 1996.
[For981 ] Forchert, Th.: "Onboard Diagnostic Systems
for Powertrain Management", European Conference on
Vehicle Electronic Systems, Proceedings, Coventry United Kingdom1998.
University Presentations
[B~ik982] Baker, B.; Forchert, Th.: " Vehicle Diagnosis
- Concepts and Tool Environment - Maintenance and
Diagnosis Systems for Mechatronic Components in
Vehicles", Center for Design Research (CDR), Dept.
Mechanical Engineering, Stanford Univ. 1998.
Dissertation or Thesis
[B~k981] B~ker, B.: "Energie- und Informationsmanagementfiir zukiinftige Kfz-Bordnetze", Technical University of Braunschweig, Dissertation 1998, Publisher
Verlag Mainz,
ISBN 3-89653-241-3,
Aachen Germany 1998.
[Mau97] Mauss, J.: "Analyse kompositionaler Modelle
durch Serien-Parallel-Stern Aggregation", University of
Hamburg, AI Dissertation 1997, Publisher Hundt GmbH
K~ln, ISBN 3-89601-183-9, Hamburg- Germany 1997.
Research and Technology i
gnosisfor FutureMe
DAIMLERCHRYSLER
Figure1: IntroductionSlide.
1. Introduction
1,1. Overview
Diagnostics
/ Motivation
ResearchandTechnology l
Headqueters
Workehop
Vehicle
Onboard
Dlegnostios
Assistan~
Vehicle
DAIMLERCHRYSLER
Figure 2: Diagnostics
3
/ Overview.
1. Introduction
/ Motivation
Researchand Technology,,,m,,
1.2. Motivation SystemDiagnosis
f Goalsfor Onboard
Diag
-~ trol a
~plex
~tem =
¯ Onboard
diagnosis
adaptability
of the
variantsandconfiguration
diversity
, exactfault information
¯ avoidance
of secondary
faults
¯ detailedinformation
of
suspected
component(s)
fault
ECUs
Detailed
information
about:
faulty(user-)functions
¯ degree
of importance
¯ actionstobe aken
intsrconnection
DAIMLERCHRYSLER
¯ O*~ml*rC,
hryd*rAa
||ker ~iNSt*~o~i#pl
l$.lLtt hh4
Figure 3: Motivation / Goals of a System Diagnosis.
1. Introduction
I Motivation
ResearchandTechnology /
1.3. DiagnosisProcedure
m,.m. DAIMLERCHRYSLER
Figure 4: Description
i) O|lmlerC,
krfderAG
Biker
,t/dLIN
St|~4m’d,ppt
1|i|,|l Stiles
of the Diagnosis Procedure.
4
2. Overview selected diagnosis methods(I)
Research and Technology
Prof. Isermann
Technical University of Darmstadt
FDI (Fault Detection and Isolation)
L
H(s)
A.s2+B.s+C
D.s2 + E.s+ F
adaptive
controlidentification
diagnosisdata
+ compac
+ dynamicprocessdescription possible
o optimizationof the processchain
- no multi-fault detection
- I
]ICPU necessary
+ dynamicprocessdescription possible
- complexalgorithms used
- no completeprocesschain
- no multi-fault detection
i
DAIMLERCHRYSLER
Figure 5: Description of Selected Diagnosis MethodsI.
2. Overviewselected diagnosis methods(11)
Research and Technology mamma
M odel-based Diagnosis
(qualitative) -Dr. Struss, Munich
M odel-based Diagnosis
(quantitative) with rodon-Tool
~
4~ developmentdata
l
mathematical,
physical
systemdescription
diagnosis-checklists
L ............................
...................................................
diagnosis-data
(qualitative database)
¯
CAN
j .........
i
..........................
.o.n..b.9.a..rd.
.......................
Jp
o dynamic
processdescriptionpossible
+ function-andcomponent-hierarchy
+ multi-faults detection
+ severityof faults
+ structureinformationof sourcesystem
o optimizationof the processchain
i DAIMLERCHRYSLER
o severityof faults
- no dynamicprocessdescription possible
+ optimizationof the processchain
o structure information of the sourcesystem
- fault-simulation necessary
- no multi-fault detection
eO=~mn+rcm,
smlir~ elk+rM~ll.m,~d++4~
lS.~z.IIIII,T
Figure 6: Description of Selected Diagnosis MethodsII.
5
2. Overview
selecteddiagnosismethods
(111)
Research and Technology l
DaimlerChrysler
MDS
DaimlerChrysler
Model-based
System
Diagnosis
designof functions and
components
CAD-structure data -~*
/_
B,,
input
configuration and----*
componentinformations ...................................................
output
FR-Dx
_n_ru-L
test
~scision
"
~
Onboard
....
diagnosisinformation:- fault component
- badfunction
- smallestswapable
unit
- fault classification
symptoms
faults
~
i .........
tree
+ modeldesignprocesssupportedwith iibraries
+ optimizationof the processchain
o non complexmain algorithm
+ suggestionof test measurement
locations
+ generationof decision trees
+ no fault simulation necessary
o dynamic process description possible
+ function end componenthierarchy
+ multi fault detection
+ severity of faults
+ structure information of the source system
+ optimization of the process chain
m DAIMLERCHRYSLER
Figure7: Descriptionof Selected DiagnosisMethods
III.
3. Process
Chain
Research and Technology
3.1. Automated
Diagnosis
DataGeneration
construction
layer
¯ math., phys.layer
~M1
~
PMechl
¯
..............
.................
!i
I!~_~
: eMech2
fct.
~
graph c
M1 M2
CO
l "M
l = co2 "M
2
O)1 I¯ M
:o
°,trx
Io,1;
M2---
(FM-Tensor)
0)
2
mechanical
clutch
(schematics
CAD-data)
modelinglayer
diagnosis
information
after
thresholdcalculation(example):
400nm< M2 < 560nm
&&n <3.500U/rain
=>M2o.k.
I
t
m DAIMLERCHRYSLER
Figure8: Automated
DiagnosisData GenerationProcess.
6
3. ProcessChain
Researchand Technology nmmim
3.2. ProcessChainDevelopment
corn ponent-diagnosis
autocode
generato~
functional description
(e.g. underMatrixXor
Matlab/Sireulink)
’~
px~
*,C
netlist tool
lid
dimensiondata
i
sourcecode
DAIMLERCHRYSLER
data of the
systemstructure
limits for
threshold
algorithms
q)O|lmh~’Ckrfnllr
Aamll~AAAIII
IIt~od,ppl
i1,1|.#hh!
Figure 9: Process Chain Development.
4. OverviewSystemDiagnosis
Research
and Technology
u
Diagnosis
Workshop
Driver
DAIMLERCHRYSLER
¯ Dimba~..k.’yd,e
AG
Nker
AAAII4P
IlmloN,ppt
I|,q|Jll ~llllO
Figure 10: OverviewSystemDiagnosis.
7
5. Summary
Research
andTechnology
m,,,,,,,
5.1. Outlook
electric I electronic-components
engineI powertrain
@
Newbrake I steering / suspensionsystems
i
interior / telematicsI
new-comm
unications
DAIMLERCHRYSLER
Figure 11: Summary
/ Outlook.
5. Summary
5.2. Overview
DiagnosticTechnologies
Researchand TechnologyHi,,,
Diagnostic Technologies
DAIMLERCHRYSLER
Figure 12: OverviewDiagnostic Technologies.