Cyber-Physical MBD For Multi-Physics Automotive Systems: Sujit S. Phatak DJ Mccune George Saikalis, Ph.D. Yasuo Sugure
Cyber-Physical MBD For Multi-Physics Automotive Systems: Sujit S. Phatak DJ Mccune George Saikalis, Ph.D. Yasuo Sugure
Cyber-Physical MBD For Multi-Physics Automotive Systems: Sujit S. Phatak DJ Mccune George Saikalis, Ph.D. Yasuo Sugure
183
OVERVIEW: The need for CPSs derives from the complexity of modern
automotive embedded systems, which can contain more than 100 ECUs,
a large ROM capacity, and a large amount of software code. Another
issue is that traditional design methodologies find it difficult to cope when
fundamental hardware problems are identified late in the development
process, during the validation phase. The MBD approach is one possible
solution to these issues and it lays the foundation for CPSs.
INTRODUCTION
THE defining feature of cyber-physical systems (CPSs)
is their close integration between physical processes
and systems on one hand and computational systems
on the other(1). CPSs integrate the dynamics of physical
processes with those of the software and network,
providing abstractions for modeling and design as well
as analysis techniques suitable for integrated systems.
In contrast to traditional embedded systems, where the
emphasis is more on operation as a standalone device,
CPSs emphasize the network of interacting systems.
Similarly, while traditional embedded systems focus on
the computational components, CPSs primarily focus
on the interfaces between the computational elements
of the system. Their scope of application covers all
sensor-based control systems, embedded systems,
and autonomous systems. These cover a very wide
range, from device-based systems such as automotive
systems, entertainment and home appliances to
integration systems such as social infrastructure,
energy, freight and transportation, aeronautical and
space applications, and healthcare, and also technical
platforms such as manufacturing systems.
The mechatronic control systems that are typically
implemented in automotive applications, such as
engine control, transmission control, throttle control,
and braking, typically involve multiple complex
physical systems with dedicated embedded controllers
that communicate with each other via a vehicle
network, such as Controller Area Network (CAN)
or FlexRay. Model-based design (MBD) is adopted
as a way of making the design process for these
complex system more efficient(2). The system design
stage integrates models of physical system behavior
(also called plant models) with controller models
to produce an abstracted system implementation.
*1 MATLAB and Simulink are registered trademarks of The MathWorks, Inc.
184
Abstraction
level
Design
Concept
Component
Validation
Car
maker
Concept
design
System
design
Component
Implementation
Product
Abstraction
level
Concept
System
validation
Tier 1
Design
Car
maker
Behavioral/
mathematical model
MATLAB
Detail model
Component
Tier 2
Unit validation
Time
Tier 1
Plant: MATLAB
Tier 2
Ctrl.: Micro./ECU model
Algorithm design
Behavior simulation
Rapid prototyping
ECU hardware
(CPU, ASIC, board)
BIOS software coding
RTOS configuration
Code generation
Time
185
Feedback
User/
driver
inputs
Plant models
from OEMs
Gasoline engine
HEV/EV system
Other physical systems
Engineering
units
Control/
soft ECU
model
MATLAB/
Simulink
Visualization/analysis
User/
driver
inputs
Engineering
units
Plant models
from OEMs
Gasoline engine
HEV/EV system
Other physical systems
Sensor models
Analog I/O
Digital I/O
Signal
conditioning
MATLAB/Simulink
Co-simulation
Feedback
Virtual CPU
simulator
Virtual CPU model
MPC/SH-2A/...
Voltage
signals Debugging Visualization of CPU
capability Program counter
Instruction cycles
Interrupts
Other...
Co-simulation
Direct one-to-one
Co-simulation bus
I/O: input/output
186
Implementation
System
Gasoline fuel
pump
PWM
Simulator
Pump
model
Simulator
Driver
circuit
I/O
signals
Response
pressure
Simulator
Inlet
valve
Simulator
I/O signals
Co-simulation bus
PWM: pulse width modulation
Battery voltage
Target pressure
MATLAB/Simulink
I/O signals
Feedback
Co-simulation bus
Control model
Simulator
MATLAB/
Simulink
Simulator
I/O signals
Target
pressure
I/O signals
Feedback
Virtual-CPUbased implementation
Simulink
implementation
Battery
voltage
Battery
voltage
Target
pressure
Virtual CPU
Device drivers,
RTOS
MATLAB/
Simulink
C auto
code
I/O signals
Target
pressure
Virtual CPU
simulator
MATLAB/
Simulink
Feedback
Co-simulation bus
Fig. 7Simulink-based CPS for Gasoline Fuel Pump.
This Simulink-based CPS uses a co-simulation bus.
SH-2A
I/O signals
Feedback
Co-simulation bus
187
Co-simulation bus
Virtual CPU
Gasoline fuel pump
SH-2A
Target
pressure
Virtual CPU
simulator
MATLAB/
Simulink
I/O signals
Simulator
Simulator
I/O signals
Feedback
Co-simulation bus
Fig. 9Virtual CPU-based CPS for Gasoline Fuel Pump.
The virtual SH-2A control model integrates with the plant model
of the gasoline fuel pump via a co-simulation bus.
250
250
200
200
150
100
Simulink
Virtual CPU simulator
50
Pressure (MPa)
Pressure (MPa)
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Simulink
Virtual CPU simulator
150
100
50
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
188
USA office
Japan office
I/O signals
Feedback
Co-simulation bus
Multiple PC/Multiple simulator/
Multiple location implementation
PC-1
PC-2
PC-3
Co-simulation bus
simulated using Simulink and later using a virtualCPU-based approach. Both approaches achieved a
good correlation. An international multi-location CPS
implementation was also tested and validated to prove
the flexible multi-location implementation of these
systems as a multi-physics robust design methodology.
REFERENCES
(1) S. Phatak et al., Cyber Physical System: A Virtual CPU
Based Mechatronic Simulation, IFAC (Sep. 2010).
(2) G. Saikalis et al., Virtual Embedded Mechatronics System,
SAE Technical Paper # 2006-01-0861.
(3) J. Sinnamon et al., Co-simulation Analysis of Transient
Response and Control of Engines with Variable Valvetrains,
SAE Technical Paper, 2007-01-1283.
(4) ChiasTek, 2009, http://www.chiastek.com
189
DJ McCune
Yasuo Sugure
George Saikalis, Ph.D.
Joined Hitachi America, Ltd. in 1990, and now works
at the Research & Development Division, Automotive
Products Research Laboratory. He is currently Vice
President engaged in many aspects of automotive
R&D. Dr. Saikalis is a member of the IEEE and SAE.