virtual vehicle systems simulation - university of...
TRANSCRIPT
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 1
Virtual Vehicle Virtual Vehicle Systems SimulationSystems Simulation
²1=_Te\Qb1``b_QSXY^BUQ\²1=_Te\Qb1``b_QSXY^BUQ\DY]U³DY]U³
G. Edzko Smid, Ph.D.G. Edzko Smid, Ph.D.
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 2
Introduction to VVSSIntroduction to VVSS
Computer technology adds tools to engineering.
Virtual reality for arcade, movie and as a design tool in engineering.
Imagine driving in a virtual realistic vehicle with full feedback.
An engineer can design and modify a subsystem, and analyze the vehicle ride performance immediately.
No knowledge or expertise is required for evaluating vehicle subsystems on full-vehicle simulation.
Implementing a vehicle ride control systems must be a drag and drop exercise.
Networked simulation environment allows for remote simulation subsystems and hardware in the loop
Modularity provides capability for distributed simulation and improved computational performance.
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 3
ContentsContents
• Objective and Applications
• Vehicle Model & Tire Model
• Modularity and Real-Time
• Implementation of UI in MATLAB/Simulink
• Conclusions
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 4
Back groundBack ground
• Many industry activities involve design and testing of automotive systems.
• ITT : ABS, TACOM : Leader-following, suspension, collision warning
• Simulation could be used to replace certain test-track experiments.
• Many simulation packages have black-box dynamics models, are expensive and are not flexible to fit any specific purpose.
• Collaborative online simulation exploit capabilities of internet
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 5
ApplicationsApplications
• Design of vehicle sub-systems (interactive)
• Analysis of vehicle sub-systems (state-access)
• Subjective validation of concept designs
• Facilitate testing of prototype systems
• Benchmark hazardous scenarios
• Education for drivers and customers
• Instrumentation for in-the-loop mechatronics studies
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 6
Laborator y SetupLaborator y Setup
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 7
Virtual Vehicle S ystem Virtual Vehicle S ystem
( )( )uxDuCxy
uxBuAxx
,
,
ς++=η++=State Space Formulation:
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 8
Vehicle Model Vehicle Model -- Free Bod yFree Bod y
Sum of Forces:
( ) windirollideflislipi
FFFFF +++∑==
,,,
4
1
Integration
( )( )kkkk
kkvkk
vvtxx
FFMtvv
+∆+=+∆+=
−−
−−
−
121
1
11
21
1
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 9
Vehicle Model Vehicle Model -- Free Bod yFree Bod y
Sum of Torques:
Angular Body Accelerations:
Roll, Pitch, Yaw :
( )irollio
ideflio
islipio
iFAFAFCT ,,,
4
1×+×+×∑=
=
( )( ) 1−Ω×Ω−=Ω IIT cccc
Ω
θφθφθ−φ
θφθφ=Θ cc
seccossecsin0
sincos0
tancostansin1
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 10
Vehicle Model Vehicle Model -- SuspensionSuspension
( )sFss
KDss
KDzsM zSSSTTTww −=δ
++δ
+σ+ 11
cos2
( )
SSw
ic
ic
SS
ioo
io
TT
Lz
ABL
BIsectBL
δ−=
−−=δ
σ−−=δ ,
Suspension and Tire deflection spoke model
Equation of motion:
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 11
Tire Dynamics ModelTire Dynamics Model
• Existing tire models are either- not related to physical parameters- based on lookup tables
- derived from velocities; requires minimum velocity
• New tire model is introduced, based on two analytical models and physical parameters.
• Fuzzy logic is used for supervisory decision on the actuation of the models, based on the state of the vehicle.
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 12
Tire Dynamics Tire Dynamics -- DugoffDugoff
µ
wheel
vehiclewheelsυ
υ−υ=
Slip coefficient:
Slip ratio:
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 13
Tire Dynamics Tire Dynamics -- DugoffDugoff
−µ
+µ
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 14
Tire Dynamics Tire Dynamics -- Stick FrictionStick Friction
Problem : Dugoff model does not incorporate stick frictionSolution : Introduce additional friction component, based on
Fuzzy logic supervisory control.
M
1
s
1
s
1
Friction
Spring
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 15
Tire Stick Friction ModelTire Stick Friction Model
( ) ( )
+
−+−
=0
1
,,
,,
yvstickyvstick
wxvstickwxvstick
St uKvD
uus
KvvD
FStick Friction Force:
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 16
Tire Model Tire Model -- Fuzzy SwitchFuzzy Switch
0λ− λvw vv −=δ
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 17
Tire Dynamics ModelTire Dynamics Model
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 18
Hardware Im plementationHardware Im plementation
Simulation Platform Integrated Network Environment (SPINE)
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 19
Modularit y Modularit y -- MultiMulti --CPUCPU
SPINE is the communication backbone for the VVSS. The simulation software for each module communicates through a DLL with the VVSS. The VVSS serves a pool of shared memory to the modules through SPINE.
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 20
Modularit y Modularit y -- State Indexin gState Indexin g
( )
( )
( )n
i
1
xL
xL
xL
ix&
ix 1~i
x 2~i
x
Implicit Bilinear Approximation:
( )
( )kkkk
kkkk
kkk
xxh
cxx
xxh
bxx
ahuxx
,21,2,31,3
,11,1,21,2
,11,1
2
2&&&
&&&
&
++=
++=
+=
++
++
+
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 21
kkkkk
kkk
kkk
uh
abcxh
bcxchxx
uh
abxbhxx
ahuxx
k
42
2~
3
,1
2
,2,31,3
2
,1,21
,11,1
1,2
+++=
++=
+=
+
+
+
&&&
&&
&
Modularit y Modularit y -- IntegrationIntegration
Implicit Bilinear Approximation:
Explicit Bilinear Approximation:
s
a
s
b
s
c
Forward Euler Approximation:
kkk
kk
kkk
xchxx
xbhxx
ahuxx
k
,2,31,3
,1,21
,11,1
1,2
~
&&
&&
&
+=
+=
+=
+
+
+
kht =At Some:u
1x12
~x 3x
( )
( )kkkk
kkk
kkk
xxh
cxx
xxh
bxx
ahuxx
k
,21,2,31,3
,11,1,21
,11,1
2
2~
1,2
&&&
&&&
&
++=
++=
+=
++
+
+
+
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 22
Modularit y Modularit y -- Timin gTimin g
of a subsystem
of a subsystem
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 23
Real time Real time -- Simulation timeSimulation time
n
ntth s
s
+> 0
fs Tntt <+0
Time to compute overhead and n integration steps must be smaller than frame size:
The time that is simulated per framemust match the real time laps of the frame:
n
Th f
s =
The minimum sub-sample integration time must meet the requirement:
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 24
Real timin gReal timin g
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 25
Implementation in MATLABImplementation in MATLAB
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 26
Implementation in MATLABImplementation in MATLAB
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 27
Implementation in MATLABImplementation in MATLAB
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 28
Implementation in MATLABImplementation in MATLAB
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 29
Real Timin gReal Timin g
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 30
Driver EnvironmentDriver Environment
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 31
ConclusionsConclusions
Developed a Vehicle Simulation Environment (VVSS) and a real-time based simulation network (SPINE).
Facilitate a plug-and-play ready to use simulation environment in familiar popular Matlab/Simulink
Fully interactive audio/motion feedback stereo-vision human driver environment
Easily configure any driving scenario and terrain configuration through separate datafile
Still unsolved problem for P&P calibration ...
AUTOMOTIVE RESEARCH CENTER
Intelligent Vehicle Dynamics and Control
arc NAC
Oakland UniversitySlide 32
Virtual Vehicle Virtual Vehicle Systems SimulationSystems Simulation
²1=_Te\Qb1``b_QSXY^BUQ\²1=_Te\Qb1``b_QSXY^BUQ\DY]UDY]U³³
Fine.Fine.