ascl of jilin ascl of jilin universityuniversity€¦ · tra ic (cars/pedestrian/signal/signs)...
TRANSCRIPT
ASCL of Jilin ASCL of Jilin UniversityUniversity吉林大学汽车仿真与控制国家重点实验室吉林大学汽车仿真与控制国家重点实验室
Confidential 11
Background & MotivationBackground & Motivation
Challenges Technological Challenges on Vehicle
Intelligence
Advancementg
ModelingModelingSimulationSimulation
Confidential
D i i i t l i t tBackground & MotivationBackground & Motivation
Driving environment now plays an important or even critical role on the safety and performance of intelligent driving Driverdriving.
DrivingDrivingweatherweather traffictraffic
ggEnvironmentEnvironment
El i C l
Confidential
Vehicle Electronic Control
roadroad
Motivation: Motivation: Challenges on Testing & ValidationChallenges on Testing & Validation
Setting up field testing for many features can be of long‐V2V Coordination
Collision Avoidance
g p g y gcycle and high‐cost, and even impossible.
Some safety critical features can be difficult, if not impossible to test in the field especially during the
Intelligent Vehicle
impossible, to test in the field, especially during the early development stages!
Confidential
V2V Coordination Pedestrian Collision
Technological AdvancementTechnological Advancement
Modeling on Battery and Electric Motors, on‐board communication
VehicleDynamics
driverdriver
weatherweather traffictrafficroadroadtiretire
radarradarwirelesswirelesscameracamera
Confidential
Modeling on Driving Environment andEnvironmental Sensing
Modeling & Simulation Plays Key Roles Modeling & Simulation Plays Key Roles
Advanced Development PlatformAdvanced Development Platform
Pure SimulationPure Simulation离线纯仿真平台离线纯仿真平台
RTRT--SIL/HIL SimulationSIL/HIL SimulationRTRT SIL/HIL SimulationSIL/HIL Simulation实时、软硬件在环仿真平台实时、软硬件在环仿真平台
DIL with 3D EnvironmentDIL with 3D Environment驾驶员在环三维场景仿真驾驶员在环三维场景仿真驾驶员在环三维场景仿真驾驶员在环三维场景仿真
InIn--Vehicle TestVehicle TestConfidential
InIn--Vehicle TestVehicle Test实验车试验实验车试验//标定标定//验证验证
Tools & Method on Vehicle Tools & Method on Vehicle IntelligenceIntelligence
Vehicle Performance
Virtual driving environment(road, traffic, weather, etc)
Virtual driving scenarios(ABS/TCS/ESP/ACC, self-parking)
l Simulated traffic
Vehicle Virtual Driving Vehicle Virtual Driving EnvironmentEnvironment
Environmental sensors(radar, camera, GPS/map, V2V, V2X)
Simulated traffic participants
(cars, pedestrians, barrels, etc)
Vehicle IntelligenceVehicle Intelligence
Confidential
Tools & Method on Vehicle Tools & Method on Vehicle IntelligenceIntelligence
Driver ModelDriver Model
Weather ModelWeather Model Traffic ModelsTraffic Models
Virtual Driving Virtual Driving EnvironmentEnvironment withwith
Environment & Sensor ModelsEnvironment & Sensor Models
Vehicle Dynamics ModelVehicle Dynamics Model Electronic Control
Confidential
Road & Field ModelRoad & Field Model
Vehicle Dynamics Modeling: Vehicle Dynamics Modeling: with Highwith High--Order, LargeOrder, Large--Nonlinearity Nonlinearity with Highwith High Order, LargeOrder, Large Nonlinearity Nonlinearity
and High Efficiencyand High Efficiency
Confidential
Vehicle Dynamics ModelVehicle Dynamics Model
RearSprung Mass
Unsprung Mass
Unsprung Mass
SystemSystem--based modeling based modeling approach with:approach with:hhighigh--order and largeorder and large--
Wheel Spin
p g gg ggnonlinearitynonlinearityhigh efficiency for realhigh efficiency for real--timetimehigh fidelityhigh fidelityp
SideslipFront
high fidelityhigh fidelity
Front
sw
~
Confidential
Vehicle Dynamics Model Vehicle Dynamics Model (under (under SimulinkSimulink))
Confidential
Vehicle Dynamics Model: Vehicle Dynamics Model: verificationverification
10
20
30
40
gle
[deg
]
CarsimPanosimDouble-Lane Change
-30
-20
-10
0
Ste
ering
Whe
el A
ng
Simulated speed at 80kp
0 1 2 3 4 5 6 7 8 9 10-40
Time [s]
Steering Angle
10
15 CarsimPanosim
0.2
0.3
0.4 CarsimPanosim
5
0
5
Yaw
Rat
e [d
eg/s
]
-0.1
0
0.1
tera
l Acc
eler
atio
n [g
]
-15
-10
-5
-0.4
-0.3
-0.2
Lat
ConfidentialLateral GYaw Rate
0 1 2 3 4 5 6 7 8 9 10-15
Time [s]0 1 2 3 4 5 6 7 8 9 10
0.4
Time [s]
Vehicle Dynamics Model: Vehicle Dynamics Model: verificationverification
Fishhook Simulation
Simulated speed at 80kph, with steering and path shown below
300 20
100
200
300
deg]
CarsimPanosim
-20
0
20CarsimPanosim
-100
0
100
erin
g W
heel
Ang
le [d
-60
-40
Y C
oord
inat
e [m
]-300
-200
-100
Ste
e
120
-100
-80
PathSteering Angle
0 1 2 3 4 5 6 7 8 9 10-300
Time [s]0 50 100 150
-120
X Coordinate [m]
Confidential
Vehicle Dynamics Model: Vehicle Dynamics Model: verificationverification
30 CarsimP i 0 6
0.8 CarsimPanosim
Fishhook Simulation
Yaw R
10
20
[deg
/s]
Panosim
0.2
0.4
0.6
erat
ion
[g]
Panosim
LateraRate
20
-10
0
Yaw
Rat
e [
-0.4
-0.2
0
Late
ral A
ccel
e
al G
0 1 2 3 4 5 6 7 8 9 10-30
-20
Time [s]
0 1 2 3 4 5 6 7 8 9 10
-0.8
-0.6
Time [s]
2.5 Carsim
80.4 Carsim
0.5
1
1.5
2
deg]
CarsimPanosim
79.6
79.8
80
80.2
ty [k
m/h
]
CarsimPanosim
Roll
1 5
-1
-0.5
0
Rol
l Ang
le [d
79
79.2
79.4Lo
ngitu
dina
l vel
ocit
l Ang
le
Speed
Confidential0 1 2 3 4 5 6 7 8 9 10
-2.5
-2
-1.5
Time [s]
0 1 2 3 4 5 6 7 8 9 10
78.4
78.6
78.8
Time [s]
Vehicle Dynamics Model: Vehicle Dynamics Model: verificationverification
Fishhook Simulation
2.5 CarSimP Si
4 CarSim
1
1.5
2
gle
[deg
]
PanoSim
1
2
3
km/h
]
PanoSim
Sides
Late
-0.5
0
0.5
Veh
icle
Slip
Ang
-1
0
1
Late
ral S
peed
[k
slip A
ngle
eral Spee
0 1 2 3 4 5 6 7 8 9 10-1.5
-1
Time [s]
0 1 2 3 4 5 6 7 8 9 10-3
-2
e edTime [s]
Confidential
Radar Modeling: Radar Modeling: with Combined Geometric and Physical with Combined Geometric and Physical with Combined Geometric and Physical with Combined Geometric and Physical
ApproachesApproaches
Confidential
Radar ModelingRadar Modeling Take into consideration of major physical factors, such as
electromagnetic wave propagation (echo signals), target and its RCS estimation and ambient noiseestimation, and ambient noise
Combined geometric and physical modeling approaches to achieve high fidelity while maintaining real-time computational efficiency
Target M d l
high fidelity while maintaining real time computational efficiency
Model
RCS Echo SignalRadar Model RCS
Estimation
gPropagation
Model
Ambient Model
Confidential
Radar ModelingRadar Modeling
Confidential
Radar ModelingRadar Modeling
Confidential
Camera ModelingCamera ModelingCamera ModelingCamera Modeling
Confidential
PanoSimPanoSim::Camera ModelingCamera Modeling
Confidential
Basic Camera Modeling
PanoSimPanoSim::Camera ModelingCamera Modeling
fr)2tan(2 fr
)2i (2 f)sin(fr
)2sin(2 fr
Fisheye & Wide Angle Camera Modeling and Validation
Fisheye Camera Modeling
Confidential
Camera Modeling and Validation
PanoSimPanoSim::Camera Mounting & CalibrationCamera Mounting & Calibration
Take PicturesTake Pictures
Record Video
Confidential Video Playback
PanoSimPanoSim::Camera Model ApplicationsCamera Model Applications
Traffic Sign Recognition Distance/Speed MeasurementTraffic Sign Recognition Distance/Speed Measurement
Confidential
Lane Marker DetectionObject Detection
Highly Correlated with the Results from Real ImageHighly Correlated with the Results from Real Image
Real Image Simulated Image
Lane detection using Hough transform
Road detection using watershed algorithm
Confidential
ddAn Integrated An Integrated Software ToolSoftware Tool
Confidential
PanoSimPanoSim::Six Modules with Many FunctionsSix Modules with Many Functions
Main GUIs VehicleBuilder
FieldBuilder PanoCam
PanoPlot/PanoAnim MDL Generator
T ffiR d /Lid /CV hi l D i
Confidential
Traffic Weather/Light
Radar/Lidar/Camera Virtual Proving Ground
Vehicle Dynamics Wireless Comm
PanoSimPanoSim::ApplicationsApplications
Vehicle Performance
Virtual Proving Ground(Road/Traffic/Weather)
Virtual Scenarios(ABS/ESP/ACC/LDW/UPA)
l ffEnvironmental Sensors(Radar/Lidar/Camera/Wireless/GPS)
Traffic(Cars/Pedestrian/Signal/Signs)
ADASADAS
Confidential
PanoSimPanoSim::ApplicationsApplications
ConceptSystem
Performance
PrototypeSubsystemIntegration
ProductSubsystem
lid iProduct Validation
An Integrated Platform for Technology and Product Development
of Vehicle Intelligence
Confidential
of Vehicle Intelligence
PanoSimPanoSim::ApplicationsApplications
Vehicle Performance
ADAS ACC/LDW/LKA/FCW Autonomous Parking
Brake/Steer/Suspension Handling & Stability
g
Environment Sensing Vision-based sensing & IP
Powertrain System Engine/Transmission Driveline & Driveability
g Radar Detection Sensor Fusion
I t lli t D i iy
Electronic Controls ABS/TCS
Intelligent Driving Planning & Decision Positioning/NavigationABS/TCS
ESPIntelligent Transportation V2V/V2I/V2X
A D i i
Confidential
Autonomous Driving
FieldBuilderFieldBuilder::Building Testing FieldBuilding Testing Field
3D Virtual Proving Ground
Confidential
TrafficTraffic::Traffic Modeling for ADASTraffic Modeling for ADAS
Good fidelity in traffic dynamics, while maintaining sufficient computational efficiency
Good flexibility in generating desired disturbances for ADAS feature development, while maintaining its nature of randomness
Confidential
PanoSimPanoSim::Virtual Testing & VerificationVirtual Testing & Verification
C lli i Collision Avoidance
Adaptive Cruise Control
Confidential
Autonomous Parking
PanoSimPanoSim::Seamless Seamless Connection to SimulinkConnection to Simulink
Auto Simulink Model Generation
Hi h Effi i t High Efficient Numerical Simulation
Confidential
RealReal--Time HIL/DIL Time HIL/DIL SimulationSimulation
Confidential
Simulation on Simulation on Vehicle Intelligence DevelopmentVehicle Intelligence Development
Extend the functionality of driver simulator with environmental sensingenvironmental sensing
Confidential
Camera Modeling: ApplicationsCamera Modeling: ApplicationsReal-time HIL/DIL simulation with camera model
Virtual Driving EnvironmentVirtual Driving Environment
Driver SimulatorCamera Model
3Cat Warmup Scalar for Costs
2Cat Warmup Scalar for Rate L imits
1Cat al yst Warmup Enabl ed
single(1)
single(0)
Action
Eng C at Warmup Enabled
Cat Warmup Scalar
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Use Cat Warmup Fl ag from Eng
Max R at e
Min Rate
Max
Min
input
output
Rate and Range Limiter 2
Max R at e
Min Rate
Max
Min
input
output
Rate and Range Limiter 1
Merge
Merge3
Merge
Merge2
Merge
Merge1
u1if (u1)
else
If
Action
Cat Warmup Scalar
Cat Warmup Threshold
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or C osts
Contro l without Flag f rom Engine
7Ti me Step
6Min Rate Li mi t for Cost Scal ar
5Min Rate L imit for Rate Scalar
4Cat Warmup T hreshold
3Use Eng Cat Warmup Fl ag
2Cat Warmup Scalar
1Eng Cat Warmup Enabl ed
VeEOCR_r_SclForR ateLimRaw
VeEOCR_r_SclCatWarmupCost Raw
VeEOCR _r_SclC at WarmupCost
VeEOCR _r_SclForRateLim
3Cat Warmup Scalar for Costs
2Cat Warmup Scalar for Rate L imits
1Cat al yst Warmup Enabl ed
single(1)
single(0)
Action
Eng C at Warmup Enabled
Cat Warmup Scalar
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Use Cat Warmup Fl ag from Eng
Max R at e
Min Rate
Max
Min
input
output
Rate and Range Limiter 2
Max R at e
Min Rate
Max
Min
input
output
Rate and Range Limiter 1
Merge
Merge3
Merge
Merge2
Merge
Merge1
u1if (u1)
else
If
Action
Cat Warmup Scalar
Cat Warmup Threshold
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or C osts
Contro l without Flag f rom Engine
7Ti me Step
6Min Rate Li mi t for Cost Scal ar
5Min Rate L imit for Rate Scalar
4Cat Warmup T hreshold
3Use Eng Cat Warmup Fl ag
2Cat Warmup Scalar
1Eng Cat Warmup Enabl ed
VeEOCR_r_SclForR ateLimRaw
VeEOCR_r_SclCatWarmupCost Raw
VeEOCR _r_SclC at WarmupCost
VeEOCR _r_SclForRateLim
Image P
ControllerRT Vehicle Model
Processor
ConfidentialHIL
行驶环境及交通模拟
传感数据
系统功能演示:车辆动力学实时仿真汽车ACC电子控制算法
虚拟传感器驾驶员动态模拟器
数据处理
实时车辆动力学模型
汽车ACC电子控制算法环境与环境传感模拟驾驶员在环动态模拟车载通信及网络模拟
控制算法
ACC控制命令
(硬件在环)
车载通信及网络模拟人机界面设计汽车电控与智能化系统集成
3Cat Warmup Scalar for Costs
2Cat Warmup Scalar for Rate L imits
1Cat al yst Warmup Enabl ed
singl e(1)
singl e(0)
Action
Eng Cat Warmup Enabled
Cat Warmup Scalar
C at Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Use Cat Warmup Flag f rom Eng
Max Rate
Min Rate
Max
Min
input
output
Rate and Range Li mi ter 2
Max Rate
Min Rate
Max
Min
input
output
Rate and Range Li mi ter 1
Merge
Merge3
Merge
Merge2
Merge
Merge1
u1if (u1)
else
If
Action
Cat Warmup Scalar
Cat Warmup Threshold
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Contro l wi thout Fl ag from Engi ne
7Time Step
6Mi n Rate L imit for Cost Scalar
5Min Rate Limit for Rate Scal ar
4Cat Warmup Threshold
3Use Eng Cat Warmup Fl ag
2Cat Warmup Scal ar
1Eng Cat Warmup Enabl ed
VeEOC R_r_SclForRateLimR aw
VeEOC R_r_SclC at WarmupCostRaw
VeEOC R_r_SclCatWarmupCost
VeEOC R_r_SclForRateLim
3Cat Warmup Scalar for Costs
2Cat Warmup Scalar for Rate L imits
1Cat al yst Warmup Enabl ed
singl e(1)
singl e(0)
Action
Eng Cat Warmup Enabled
Cat Warmup Scalar
C at Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Use Cat Warmup Flag f rom Eng
Max Rate
Min Rate
Max
Min
input
output
Rate and Range Li mi ter 2
Max Rate
Min Rate
Max
Min
input
output
Rate and Range Li mi ter 1
Merge
Merge3
Merge
Merge2
Merge
Merge1
u1if (u1)
else
If
Action
Cat Warmup Scalar
Cat Warmup Threshold
Cat Warmup Enable
Scalar f or Rate Limits
Scalar f or Costs
Contro l wi thout Fl ag from Engi ne
7Time Step
6Mi n Rate L imit for Cost Scalar
5Min Rate Limit for Rate Scal ar
4Cat Warmup Threshold
3Use Eng Cat Warmup Fl ag
2Cat Warmup Scal ar
1Eng Cat Warmup Enabl ed
VeEOC R_r_SclForRateLimR aw
VeEOC R_r_SclC at WarmupCostRaw
VeEOC R_r_SclCatWarmupCost
VeEOC R_r_SclForRateLim
目标检测道路识别
人机界面(HMI)
ACC警示信息
Confidential
人机界面(HMI)
原型控制器
DriverDriverSimulatorSimulator
Engine HIL Engine HIL PlatformPlatform
Steering HIL Steering HIL
Environment Environment and Sensor and Sensor
ModelsModelsggPlatformPlatform
Driveline HIL Driveline HIL
Braking HIL Braking HIL PlatformPlatform
PlatformPlatform
Electric DriveElectric DriveElectric Drive Electric Drive HIL HIL PltformPltform Powertrain Powertrain
HIL PlatformHIL Platform
Confidential
Virtual Driving EnvironmentVirtual camera, radar, V2VVirtual camera, radar, V2V
Traffic and 3D environment
DIL SimulatorDIL simulation
HMI
P C llRT Vehicle Controls
Confidential
Prototype ControllerBy-wire actuators
In-vehicle communications
Conclusions:
Field testing can be very challenging on cost, time, d fl ibili d i i ibland flexibility, and sometimes even impossible.
Virtual driving environment can be in high fidelity g g ywith models of road, traffic and weather, and environmental sensors of radar and camera (and (V2V next).
It is proved to be very effective in the early design It is proved to be very effective in the early design, test and verification of vehicle intelligence.
Confidential