1、1 Public| ETAS/ESC| 2013-06-15| ETAS GmbH 2013. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.Virtual Environment for ADAS HiL TestConsulting and Engineering/Mr. Le YangConfi
2、dentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.2MotivationToday release testing of todays ADAS with up to 2 million test km an
3、d 1.000 test driversTomorrow Increasing system complexity of future ADAS will increase diversity of relevant test scenarios Costs for such release testing: several 100 million EURSolution: Looking for sustainable and affordable concept for test and validation on ADAS systemsystem Activation frequenc
4、y per 10.000km Km till activationDistance warning 40-60 170-250Breaking assistance BAS Plus 0.5-1 10.000-20.000PRE-SAFE breaking level1 0.1-0.2 50.000-100.000PRE-SAFE breaking level2 0 -Data basis:2.000.000km, 1000 drivers; source: Dr. Markus Fach et al., Daimler AG, VDI/VW Gemeinshafts-Tagung, 2010
5、Source: Prof. Winner et al., Darmstaedter Kolloquium “Mensch und Fahrzeug”, 2011Confidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property r
6、ights.4Motivation“These events are rare and our engineers carefully study these simulated contacts and refine the software to ensure the self-driving car performs safely,“ wrote Google in the report. A software “fix“ is tested against many miles of simulated driving then tested again on the road, th
7、e company explained. When the fix has been reviewed successfully, it is then rolled out to the entire fleet.money Cost reductiontime Saving time for test preparationrisk High coverage of critical testsreuse Repeatable and traceableEach time a test driver takes over from the self-driving car, were ab
8、le to play back the exact situation and predict via simulationwhat could have happened if the car had been left to drive itself.source: Googles self-driving car monthly report, 2015GooglesCars Drive ThreeMillionSimulatedMiles Every DayConfidentiality | Department | Date | ETAS GmbH 2012. All rights
9、reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.6SensorModelRequirements on ADAS HiL TestI/OinterfaceManeuverControlRealtimePCExpertise databaseDriving scenarioSensor model Maneuver contr
10、ol Realtime PCI/OScenariodatabaseIncluding the test cases from real and dummy tests, but not limited to them!ETASEngineering Service and KnowhowConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, a
11、s well as in the event of applications for industrial property rights.7Requirements on ADAS HiL TestRealtime PCSensor modelI/O interfaceManeuver controlScenario database High performance for calculation Multi core processor Open for third party model integration Open architecture for customized I/O
12、CAN, LIN Ethernet ADC, DAC, PWM and etc. Third party I/O for extension Camera/light model Environment filter for the rain, fog, night lights and etc. Camera filter for blur, glow effects, noises, lens distortion, depth of filed. Ultrasonic model Radar Model etc. Mandatory function test-e.g. AEB-City
13、 Market Standard function test-e.g. ACC. Light Differentiating function test-e.g. Traffic Jam Assist Open-loop driver input Course following-e.g. double lane change Virtual driver-different styles etc.Confidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any dispo
14、sal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.8ADAS HiL Solution-project example for Camera based systemDA sensorproviding object information for validation of objects tracking and tracingGround truth sensor for CameraD
15、epth cameraproviding deepness values for validation of stereocamera algorithmsLane marker sensorproviding lane markerinformation for validation of LKA or LDWGround truth data is used for calibration of the function under testConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved,
16、also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.9 Sequence Generation Sequence check and store (Depositing of sequences on the file server) Property Store search for Sequences Sequence labeling for
17、 measurement programs and functions HIL run for performance evaluationADAS HiL Solution-project example for Camera based systemCurrent solution for validation of camera based systemOpen loop HiLConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, ex
18、ploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.10ADAS HiL Solution-project example for Camera based systemTestfleetVideo dataCANlogsPedestrian1,100m; car,110m12445.2smeasurement databasePerception validationControl validationre
19、playRecorded data Manually labled Ground truthHundreds of hours in boring workLane, day19341.3sPedestrian3, 210m; car, 120m26343.7sVehicle CANMDF Camera calibrationschychroniztingOnly open loopis possible! Drive Millions of kilometersConfidentiality | Department | Date | ETAS GmbH 2012. All rights r
20、eserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.11ADAS HiL Solution-project example for Camera based systemTestfleetVideo dataCANlogsPedestrian1,100m, car;110m12445.2smeasurement databaseP
21、erception validationControl validationreplayRecorded data Manually labled Ground truthLane, day19341.3sPedestrian3, 210m; car, 120m26343.7sVehicle CANMDF Camera calibrationschychroniztingrisk timemoneyOnly open loopis possible! Drive Millions of kilometersHundreds of hours in boring workConfidential
22、ity | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.12ADAS HiL Solution-project example for Camera based systemsimu.fleetVideo dataCANlogs
23、Pedestrian1,100m; car,110m12445.2sAutomatically generatedPerception validationControl validationRealtimesystemsimulated data Simulated Ground truthIn mini. secondsLane, day19341.3sPedestrian3, 210m; car, 120m26343.7sVehicle CANMDF Camera calibrationschychroniztingClose loop Online virtual drivingCon
24、fidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.13ADAS HiL Solution-project example for USS based systemConfiguration example
25、s Ultrasonic Sensor USSUSS Communication S/E to ECU USS Echo-pulse effectConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.1
26、4ADAS HiL Solution-project example for USS based systemUSS model configuration Mount position of USS Detection Range and radiation pattern of USS Aiming direction of each USSDA sensor is used for the generation of ground truth informationUSS model output Index of activating USS. Detected distance to
27、 object Detected speed of object Detected theta and phi of objectUSS model output can be customized according to different requirements In current project, only detected distance&time of objects to USS are implemented. With the moving ego vehicle, the detected distance&time will be updated each 10 m
28、s.USS model implementationConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.15ADAS HiL Solution-project example for USS base
29、d systemI / OU S S A P A *U S S * S W f u n c t i o nS / EF P G Ac o r eS / ER e a l t i m e P Cs i m u l a t i o nM o d e lP u b l i c C A N M s g sU S BD i s t a n c eE c h o t i m eU S S i n t e r f a c eE m u l a t o rC u r r e n tS i m u l a t i o nF a i l u r eS i m u l a t i o nR e a l o b j
30、e c t sC A Ndifferent test level Manually open loop test Predefined distance Predefined echo time Hardware in the loop test(HiL) Predefined scenarios Virtual vehicle Virtual sensor Vehicle in the loop test(ViL) Predefined scenarios Real vehicle Virtual sensorSystem architecture: APA in the Loop syst
31、em for function testUSS*:Ultrasonic sensorAPA*:Active Park AssistConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.16ADAS Hi
32、L Solution-project example for MRR based systemConfiguration examplesFunction examples 77GHz Detection range: 0-160m Field of view: 45 Angular accuracy:+/-0.3 Cycle time: 60ms Optional: sensor Data fusion with Video ACC: Adaptive Cruise Control AEB: Automatic Emergency Braking EBA: Emergency Brake A
33、ssist PCW: Predictive Collision Warning etc.Confidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.17ADAS HiL Solution-project ex
34、ample for MRR based systemSimple MRR model Object is scanned by ideal propagation path The relative speed, angle, distance of the target objects to sensor are calculated and transmitted to the model interface.DA sensor is used for the generation of ground truth information of the target objectsExten
35、ded MRR modelFollowing effects will be considered Radar Cross-Section The absorb and scatter effect Energy loss during propagationExtended MRR model is used for the validation of signal processing algorithms in MRR Model interface is defined by the ECU supplier With the simplified MRR model, the con
36、trol algorithms can be validated.MRR model implementationConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.18ADAS HiL Soluti
37、on-project example for MRR based systemSW adaption in MRR Signal processing is bypassed software adaptation for TOI SDU need to be masked TOI data is updated each 40 ms TOI data is calculated by DA sensorValidation of SW function Activation conditions of functions Operation prevention conditions Fun
38、ctional reaction to situationsM R R *S W f u n c t i o nR e a l t i m e P Cs i m u l a t i o nM o d e lP u b l i c C A N M s g sC A ND i a g n o s t i cC A Ns i g n a l p r o c e s s i n gT O IanttenaC A NF a u l t M a s k i n gSystem architecture: MRR in the Loop system for function testMRR is the
39、control unit with the sensor integrated, the signal processing part need to be bypassed for simulation data injectionMRR*:Middle Range RadarTOI*: Target/Object InjectionSDU*:Sensor diagnostic UnitConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal,
40、exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.20ADAS Network HiLData fusion of different sensors is rapidly increasing in ADAS domain.-fusion validation testinteraction between different ADAS ECUsrequires extended test matri
41、x.-integration test Actuator ECUs perform the command of ADAS ECUs. -reaction chain testOur Mission We develop and market innovative connected devices and tailor-made solutions for high efficient automation test tool. Our competencies in electronics, sensor technology and software enable new archite
42、cture of ADAS network HiLOur Products We supply scalable Labcar system and engineering expertise designed to make everyday test intelligent, thus enable cost down, money saving and risk reduction for ADAS development and test.Customer requirementsECU close loopCustomized test systemTest ServiceETAS
43、products and solutions for the Network HiL systemExperiment Environment Test Automation ToolReal time SystemI/O interfaceADAS ECUs extension Act. ECUsConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribut
44、ion, as well as in the event of applications for industrial property rights.21ADAS Network HiLH o s t P C M P Cc a m e r aL D W f u n c t i o n sL K A f u n c t i o n sV D I *A D E *S f t . b y p a s sL a b c a rT S MA C C S WF u n c t i o nS i g n a l p r o c e s s i n gT O I *S f t . b y p a s sH
45、A S IA P AU S SS WF u n c t i o nD / AP r i v a t e p r o t o c o lS f t .E S PP u b l i c C A N R e a l n o d eR e a l n o d eR e s t B U SP r i v a t e C A NC A N o b j .R e a l N o d eD i s t a n c e & e c h oC A N C a r dU S BE t h e r n e t p o r tC A N C a r dV i d e o d a t a s t r e a mS e n
46、 s o r M o d e l *S f t .E M SS f t .E P SS f t .T C UR e s t B U SU S S M R Ro b j e c t sD r i v i n g s c e n a r i oC a r s i mE t h e r n e t p o r tH a r d w a r e i n t e r f a c eS i m u l a t i o n m o d e ltest level fusion validation test integration test reaction chain test More possibil
47、ities can be inspirited along with the development of technologiesConfidentiality | Department | Date | ETAS GmbH 2012. All rights reserved, also regarding any disposal, exploitation, reproduction, editing, distribution, as well as in the event of applications for industrial property rights.22Questions?
