SlideShare a Scribd company logo
1 of 28
Download to read offline
Event Data Recorder (EDR)
for Self Driving Vehicles
Dr. Manish Kochhal, Ph.D.
Staff Software Engineer, Autonomous Drive Group
NIO Inc
opinions expressed are solely my own and do not express the views or opinions of my employer.
Outline of the
talk
What is an EDR?
Motivations
Implications of Self-Driving Cars on EDR
Capturing Data and
Decision Making
Sensor Fusion – Cameras, Radar, LiDAR,
Automatic Control - Drive-by-wire, Brake-
by-wire, Steering Assist
EDR Implementation – An Embedded and Distributed
Real-time System Design Challenge!
Standards for EDR format, tools etc.
opinions expressed are solely my own and do not express the views or opinions of my employer.
What is an EDR?
• EDR is a tiny blackbox embedded by automakers in the
airbag safety systems to record data around the time of
the accident.
• EDR data is usually kept secret and shared with NHTSA
engineers for accident reconstruction, if an
investigation occurs.
• EDR can also be analyzed for understanding crash
dynamics and improving vehicle safety.
• EDR blackboxes are not actually mandated.
• Privacy concerns have caused the NHTSA to advise
automakers only to capture tiny bits of information.
Image: autobody-review.com
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR – A causal record of events
around the time of the accident
Image: squarell.com
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR Data – In Non-AV Cars
Steering Wheel Angle - Steering wheel position, from 0º if wheel is in the neutral
position, to +/- 1080º if wheel is turned 3 times to either direction.
Lateral Acceleration - A measure of how hard the vehicle is cornering. Lateral
acceleration can also be impact induced.
Longitudinal Acceleration - A measure of how hard the vehicle is accelerating or
decelerating. Longitudinal acceleration can also be impact induced.
Yaw Rate - A measure of the rate at which the vehicle is changing its degree of
rotation. It is important to note that this measurement is a rate.
Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert.
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR Data - Reports
Roll Rate - A measure of the rate at which the vehicle is changing its degree of roll. It
is important to note that this measurement is a rate.
Delta-V - The change in velocity is calculated from accelerometer data in the Airbag Control Module and presented
for longitudinal and lateral directions. In order to determine actual Delta-V for the crash, engineers utilize both
datasets to determine the principle direction of force for the crash.
Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert.
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR Data – Countdown to crash
Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert.
opinions expressed are solely my own and do not express the views or opinions of my employer.
Sensors Publish event data to EDR over CAN
Source: Event Data Recorder (EDR) Position Statement, Association for the Advancement of Automotive Medicine
Design of ARM based Enhanced Event Data Recorder &Evidence Collecting System, N. P. Sirsikar, P. H. Chandankhede, IOSR-JECE, Volume 9, Issue 5, Ver. V (Sep - Oct. 2014)
opinions expressed are solely my own and do not express the views or opinions of my employer.
Levels of Vehicle Autonomy
Image: Automated Driving Systems, A Vision for Safety, NHTSA
opinions expressed are solely my own and do not express the views or opinions of my employer.
ASIL - Automotive Safety
Integrity Levels
• 5-level scale (QM, A, B, C, D)
• QM means, “standard Quality Assurance is
sufficient”
• From ASIL A onwards, additional risk reduction
actions must be taken
• ASIL D describes the highest risk potential.
• ASIL has requirements allocated to it.
• The defined safety goals are the top level safety
requirements (on vehicle level!)
opinions expressed are solely my own and do not express the views or opinions of my employer.
Source: https://semiwiki.com/eda/synopsys/7480-functional-safety-is-a-driving-topic-for-iso-26262/
NHTSA’s Guideline on Safety
for Autonomous Vehicles
• NHTSA has some requirements for AEB, LDW,
LKA etc. for L2+ autonomous cars.
• Following are still evolving:
• Federal and State Local Laws
• IEEE J3016 has defined what
Operational Design Domain (ODD)
means – however no accepted standard
as to what the ODDs are.
• Data Recording standards are defined
for L3, L4, and L5.
• EDR/Data Recording is important
• Assessing car/driver faults under
several ODD scenarios
• Fail to fallback or takeover by driver or
car or both.
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR of Sensing Data and
Decisions
Source: Delphi
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR needs to capture all
decisions
opinions expressed are solely my own and do not express the views or opinions of my employer.
A screenshot of what the Google car
sees approaching a right turn; inset,
the view from inside the car.
• The baseline image is the detailed area map in
grayscale.
• Layered atop that are objects identified by the
car's sensors, depicted in colorful geometric
boxes: purple for vehicles, red for cyclists,
yellow for pedestrians.
• The red and green ladders are objects that have
an immediate impact on the car's speed; in this
case, though the traffic light is green,
pedestrians prevent a turn, as does a cyclist
coming up on the right — in a spot a human
driver might easily miss.
• The flat green line shows the car's planned
route.
EDR needs to capture Path Planning,
perceived and predicted obstacles, lane
markers, construction zones, traffic signs
etc. in real-time.
opinions expressed are solely my own and do not express the views or opinions of my employer.
Predicting Bicyclist Intentions
• This is an example of a Cruise car trying to
predict the intentions of a bicyclist nearby.
• Such scenarios involve crunching a lot of
sensory inputs for a making real-time
conservative decision and accurate
predictions.
• The data, decisions, and predictions need to
be recorded in real-time in the EDR.
Complicated prediction of obstacles and
their movements need to be recorded.
opinions expressed are solely my own and do not express the views or opinions of my employer.
Advanced
BlackBox
Should be as advanced as avionics
flight data recorders.
Should be open, including the data
collected, so the truth of self driving
limitations is always available to make
the next car even safer.
Source: https://www.wardsauto.com/industry-voices/level-5-autonomy-it-ai-or-just-deep-programming
opinions expressed are solely my own and do not express the views or opinions of my employer.
Off-the-shelf Event Data
Recorders
• 5 Radars
• 4 Surround Cameras + 1 Trifocal Camera
• 1 LiDAR
• 4 Ultrasonic Sensor
• 1 IMU
• 1 GPS
• HD Maps
EDR: Capture data from 16+ sensors
opinions expressed are solely my own and do not express the views or opinions of my employer.
Design Components Involved
for EDR Development
• Mechanical design of EDR BlackBox
• E/E topology
• Distributed Communications Middleware –
clients publish interesting event data.
• Automotive Grade OS and Virtualization – to
make applications (including Data Recorder)
failsafe, redundant and reliable.
• Automotive Grade File System – data integrity
vital during read/writes.
• Wireless – upload to cloud
• Security – EDR cannot be hacked by
unauthorized users.
• Backup Power – available to save 10-15s of
post-crash EDR record
• Tools to extract EDR.
Automotive OS Features: Containment, Virtualization, Redundancy, Failover
opinions expressed are solely my own and do not express the views or opinions of my employer.
The Vehicle is a Distributed System
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR
Framework –
System
Requirements
•Urban city driving – unprotected left turns, pedestrian crossing
•Highway driving – lane detection, lane change,
•L2/L3+ features: AEB, ACC, LKAS, LDW, Assisted Parking, …
•AD Engagement, Disengagement, Decision Uncertainty, Driver
Attention, etc …
ODD Behavior Specification – specially for L4 and
L5 autonomy.
•A separate ASIL-D controller
•Supports high throughput network connectivity, 1Gbps.
•Traffic prioritization – Virtual LAN, flow prioritization based QoS
tags
•Safety Core in the EDR ECU – to support failover of recording
application
•High Throughput Read/Writes to in-memory records (Circular
Buffers)
•High Throughput Read/Writes to ASIL-D filesystem on primary and
backup storage arrays.
•PCIe or USB – automotive grade driver and stack to interface with
memory and storage.
•Storage and Memory Requirements to accommodate 16+ sensors
recording at varying rates for 10-15 seconds of pre, at, and post
accident events.
EDR ECU
opinions expressed are solely my own and do not express the views or opinions of my employer.
ODD Behavior – Classify Events severities: Normal,
Medium, High, Critical, Warning, Fault (its Context
Sensitive!)
ODD
Driver
behavior
Car behavior Environmental
behavior
Driver
engaged
Driver
Disengaged
Driver Assist
Driver
Manual
Driver Asleep AD Engaged
AD
Disengaged
Hands-OnHands-Off
Navigation
Change
Lane Status
Pedestrian
Objects
Detected
Potential
Obstacles
TSR - Status
Driver
Fallback
Object
Unclear
Safety
Decision
opinions expressed are solely my own and do not express the views or opinions of my employer.
Hypothetical ODD Event Scenario
• City Driving Scenario (Most Unpredictable!)
• AD Engaged –Normal Event
• Driver State, asleep - High
• Right turn from start to 2-way street - Normal Event
• Traffic Sign Recognition (TSR) event – STOP or Signal – Medium
• 4-way stop - High
• Decision to take an unprotected Left turn – Critical
• Unable to take a left turn – Fault
• Able to merge to a lane – Medium
• Automatic Emergency Braking – High
• Enter Construction Zone/Slow Speed – Medium
• Exit Construction Zone/Increase Speed – High
• Estimated Unprotected Pedestrian Crossing – Critical
• Enter Parking Lot – Medium
• Park – Normal
• Crash or Safety Stop or AD Disengaged or Driver Takeover – Critical
• Airbag Deployed – Fault
• Obstacle Classification – Estimate %age – Low, High, Medium, Critical
• Decision Certainty – Conservative, otherwise - Low, High, Medium, Critical
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR: Event Abstraction Framework
Event Clients Interface
(Register, Buffer, Trigger, Cancel, Trigger Status)
LiDAR Event
Client
Camera
Event Client
Navigation
Event Client
Radar Event
Client
Ultrasonic
Event Client
Driver State
Event Client
Decision Making
Event Client
Distributed Communications Middleware
Object Detection
Event Client
Event and its Attributes
(Priority, Name, Client Id, State)
Event
Context
Airbag Deployment
Event Client
AEB LDW ACC
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR: In-Memory Recording Framework, Circular
Buffering to our Rescue!
EDR
MCU
Decision
Camera
GPS/Maps
LiDAR
Radar
Gateway(CAN/ETH)
Pre
Post
Pre
Pre
At
PostPost
Pre
CB Organization
Primary CBs Secondary CBs
Tertiary CBs
opinions expressed are solely my own and do not express the views or opinions of my employer.
EDR – Blackbox storage on Distributed and
Parallel Filesystems
• TBs of data depending upon ODD
duration and complexity.
• We need highly reliable file-system.
This is a big data problem and there
are existing solutions in data center
space.
Source: Flash Memory in the emerging age of autonomy, S. Heinrich, Lucid Motors.
opinions expressed are solely my own and do not express the views or opinions of my employer.
Future Work: EDR Format,
Extraction and Visualization Tools
• Virtual Reality based Accident/Crash Reconstruction based on
BlackBox data pulled from the EDR.
• Pull the EDR from specific tools provided by the manufacturer.
• Convert it to a VR movie – again using the tools provided.
Image Source: Courtesy Google Images – Accident Reconstruction
opinions expressed are solely my own and do not express the views or opinions of my employer.
Existing
Questions in
Public Domain
In a fully autonomous car accident, the EDR data
may be the only operational driving record. Can
we trust the relationship between automakers
and investigators to allow the truth to be
revealed? How will we know what the car saw
before a crash with a human-driven car or
another bicyclist?
Currently, there is no independent infrastructure
established with guidelines in place to assess the
completeness of EDR data. However, there is
agreement for L2/L3 like driver assist features.
opinions expressed are solely my own and do not express the views or opinions of my employer.
Conclusions
General acceptable EDR content and format is a WIP
Need to get agreement from lawmakers, automotive
manufacturers, suppliers, for the EDR content
relevant to an ODD.
Customers need to be informed about the importance
of EDR.
EDR Events need to be specified for each of
autonomous driving ODD domains.
Similar to crash testing, guidelines should be set for
applicable ODDs and the EDR recorded. The data
recorded is validated.
opinions expressed are solely my own and do not express the views or opinions of my employer.

More Related Content

What's hot

UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLES
UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLESUPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLES
UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLESiQHub
 
How to Apply Functional Safety to Autosar ECU's
How to Apply Functional Safety to Autosar ECU'sHow to Apply Functional Safety to Autosar ECU's
How to Apply Functional Safety to Autosar ECU'sRenesas America
 
Automated Driver Fatigue Detection
Automated Driver Fatigue DetectionAutomated Driver Fatigue Detection
Automated Driver Fatigue DetectionArman Hossain
 
Self driving car
Self driving carSelf driving car
Self driving carzebatasneem
 
Artificial Passenger Fair
Artificial Passenger FairArtificial Passenger Fair
Artificial Passenger FairDIYA NAMBIAR
 
Autosar-software-component_0hg.pptx
Autosar-software-component_0hg.pptxAutosar-software-component_0hg.pptx
Autosar-software-component_0hg.pptxfallleaf1104
 
In‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardIn‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardRealTime-at-Work (RTaW)
 
Self-driving cars are here
Self-driving cars are hereSelf-driving cars are here
Self-driving cars are hereAndrew Ng
 
Automotive bus technologies
Automotive bus technologiesAutomotive bus technologies
Automotive bus technologiesRadwa Tarek
 
Intelligent automotive safety system
Intelligent automotive safety systemIntelligent automotive safety system
Intelligent automotive safety systemA Reddy
 
Driving behavior for ADAS and Autonomous Driving
Driving behavior for ADAS and Autonomous DrivingDriving behavior for ADAS and Autonomous Driving
Driving behavior for ADAS and Autonomous DrivingYu Huang
 
Seminar on Advanced Driver Assistance Systems (ADAS).pptx
Seminar on Advanced Driver Assistance Systems (ADAS).pptxSeminar on Advanced Driver Assistance Systems (ADAS).pptx
Seminar on Advanced Driver Assistance Systems (ADAS).pptxMohit Nayal
 

What's hot (20)

UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLES
UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLESUPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLES
UPCOMING REGULATIONS AND SAFETY STANDARDS FOR AUTONOMOUS VEHICLES
 
Sensors In Automobiles
Sensors In AutomobilesSensors In Automobiles
Sensors In Automobiles
 
Embedded C - Lecture 1
Embedded C - Lecture 1Embedded C - Lecture 1
Embedded C - Lecture 1
 
Led Matrix Headlights
Led Matrix HeadlightsLed Matrix Headlights
Led Matrix Headlights
 
How to Apply Functional Safety to Autosar ECU's
How to Apply Functional Safety to Autosar ECU'sHow to Apply Functional Safety to Autosar ECU's
How to Apply Functional Safety to Autosar ECU's
 
Automated Driver Fatigue Detection
Automated Driver Fatigue DetectionAutomated Driver Fatigue Detection
Automated Driver Fatigue Detection
 
Autonomous car
Autonomous carAutonomous car
Autonomous car
 
Self driving car
Self driving carSelf driving car
Self driving car
 
KWP 2000 and UDS Protocols | Analysis & Comparison
KWP 2000 and UDS Protocols | Analysis & ComparisonKWP 2000 and UDS Protocols | Analysis & Comparison
KWP 2000 and UDS Protocols | Analysis & Comparison
 
Smartphone based ADAS
Smartphone based ADASSmartphone based ADAS
Smartphone based ADAS
 
Artificial Passenger Fair
Artificial Passenger FairArtificial Passenger Fair
Artificial Passenger Fair
 
Autosar-software-component_0hg.pptx
Autosar-software-component_0hg.pptxAutosar-software-component_0hg.pptx
Autosar-software-component_0hg.pptx
 
In‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look ForwardIn‐Vehicle Networking: a Survey and Look Forward
In‐Vehicle Networking: a Survey and Look Forward
 
Self-driving cars are here
Self-driving cars are hereSelf-driving cars are here
Self-driving cars are here
 
Automotive bus technologies
Automotive bus technologiesAutomotive bus technologies
Automotive bus technologies
 
Artificial passenger
Artificial passengerArtificial passenger
Artificial passenger
 
Intelligent automotive safety system
Intelligent automotive safety systemIntelligent automotive safety system
Intelligent automotive safety system
 
Driving behavior for ADAS and Autonomous Driving
Driving behavior for ADAS and Autonomous DrivingDriving behavior for ADAS and Autonomous Driving
Driving behavior for ADAS and Autonomous Driving
 
Embedded Systems in Automotive
Embedded Systems in Automotive Embedded Systems in Automotive
Embedded Systems in Automotive
 
Seminar on Advanced Driver Assistance Systems (ADAS).pptx
Seminar on Advanced Driver Assistance Systems (ADAS).pptxSeminar on Advanced Driver Assistance Systems (ADAS).pptx
Seminar on Advanced Driver Assistance Systems (ADAS).pptx
 

Similar to Event Data Recorder (EDR) for Self Driving Vehicles

Unified edr for a self driving car
Unified edr for a self driving carUnified edr for a self driving car
Unified edr for a self driving carlandyaBhai
 
Automatic number plate recognition
Automatic number plate recognitionAutomatic number plate recognition
Automatic number plate recognitionSaifullah Malik
 
Requirements for the Next-Generation Autonomous Vehicle Ecosystem
Requirements for the Next-Generation Autonomous Vehicle EcosystemRequirements for the Next-Generation Autonomous Vehicle Ecosystem
Requirements for the Next-Generation Autonomous Vehicle EcosystemM. Ilhan Akbas
 
Introduction to Smartanalysis - Compliance made easy
Introduction to Smartanalysis - Compliance made easy Introduction to Smartanalysis - Compliance made easy
Introduction to Smartanalysis - Compliance made easy Descartes Systems UK Ltd
 
Data Annotation_Cars.pptx
Data Annotation_Cars.pptxData Annotation_Cars.pptx
Data Annotation_Cars.pptxssuserfb92ae
 
B - NTSB SAE EDR Symposium 2003
B - NTSB SAE EDR Symposium 2003B - NTSB SAE EDR Symposium 2003
B - NTSB SAE EDR Symposium 2003Robert McElroy
 
Career Opportunities in Electric vehicles and Autonomous vehicles .pptx
Career Opportunities in Electric vehicles and Autonomous vehicles .pptxCareer Opportunities in Electric vehicles and Autonomous vehicles .pptx
Career Opportunities in Electric vehicles and Autonomous vehicles .pptxMurali krishna U
 
Presentation - Vision
Presentation - VisionPresentation - Vision
Presentation - VisionJaweria Khan
 
Prototyping presentation
Prototyping presentationPrototyping presentation
Prototyping presentationIbtihaj Uddin
 
Atla1717 og
Atla1717 ogAtla1717 og
Atla1717 og777117
 
Driver Drowsiness.pptx
Driver Drowsiness.pptxDriver Drowsiness.pptx
Driver Drowsiness.pptxIsi India
 
automation.pptx
automation.pptxautomation.pptx
automation.pptxSabarDasal
 
Driverless Cars
Driverless CarsDriverless Cars
Driverless Carsdani007007
 
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016David Schnerch
 
How to Collect Crash Data From CARE
How to Collect Crash Data From CARE How to Collect Crash Data From CARE
How to Collect Crash Data From CARE Fatemeh Baratian
 
IRJET- Drivers Stupor Scrutinizing System
IRJET-  	  Drivers Stupor Scrutinizing SystemIRJET-  	  Drivers Stupor Scrutinizing System
IRJET- Drivers Stupor Scrutinizing SystemIRJET Journal
 
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon BaradMid-America Transportation Center
 

Similar to Event Data Recorder (EDR) for Self Driving Vehicles (20)

Unified edr for a self driving car
Unified edr for a self driving carUnified edr for a self driving car
Unified edr for a self driving car
 
Automatic number plate recognition
Automatic number plate recognitionAutomatic number plate recognition
Automatic number plate recognition
 
Requirements for the Next-Generation Autonomous Vehicle Ecosystem
Requirements for the Next-Generation Autonomous Vehicle EcosystemRequirements for the Next-Generation Autonomous Vehicle Ecosystem
Requirements for the Next-Generation Autonomous Vehicle Ecosystem
 
AUTONOMOUS VEHICLES 2.pdf
AUTONOMOUS VEHICLES 2.pdfAUTONOMOUS VEHICLES 2.pdf
AUTONOMOUS VEHICLES 2.pdf
 
Introduction to Smartanalysis - Compliance made easy
Introduction to Smartanalysis - Compliance made easy Introduction to Smartanalysis - Compliance made easy
Introduction to Smartanalysis - Compliance made easy
 
Data Annotation_Cars.pptx
Data Annotation_Cars.pptxData Annotation_Cars.pptx
Data Annotation_Cars.pptx
 
B - NTSB SAE EDR Symposium 2003
B - NTSB SAE EDR Symposium 2003B - NTSB SAE EDR Symposium 2003
B - NTSB SAE EDR Symposium 2003
 
Career Opportunities in Electric vehicles and Autonomous vehicles .pptx
Career Opportunities in Electric vehicles and Autonomous vehicles .pptxCareer Opportunities in Electric vehicles and Autonomous vehicles .pptx
Career Opportunities in Electric vehicles and Autonomous vehicles .pptx
 
Automated vehicle
Automated vehicleAutomated vehicle
Automated vehicle
 
Presentation - Vision
Presentation - VisionPresentation - Vision
Presentation - Vision
 
Prototyping presentation
Prototyping presentationPrototyping presentation
Prototyping presentation
 
Atla1717 og
Atla1717 ogAtla1717 og
Atla1717 og
 
Driver Drowsiness.pptx
Driver Drowsiness.pptxDriver Drowsiness.pptx
Driver Drowsiness.pptx
 
automation.pptx
automation.pptxautomation.pptx
automation.pptx
 
Driverless Cars
Driverless CarsDriverless Cars
Driverless Cars
 
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016
HOS-ELD_Overview_of_Features_and_Benefits_-_MDS_2016
 
How to Collect Crash Data From CARE
How to Collect Crash Data From CARE How to Collect Crash Data From CARE
How to Collect Crash Data From CARE
 
IRJET- Drivers Stupor Scrutinizing System
IRJET-  	  Drivers Stupor Scrutinizing SystemIRJET-  	  Drivers Stupor Scrutinizing System
IRJET- Drivers Stupor Scrutinizing System
 
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad
2020 MATC Summer Seminar Series: Mr. Jason Marks & Mr. Jon Barad
 
A Proposed Accident Preventive Model For Smart Vehicles
A Proposed Accident Preventive Model For Smart VehiclesA Proposed Accident Preventive Model For Smart Vehicles
A Proposed Accident Preventive Model For Smart Vehicles
 

Recently uploaded

Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...
Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...
Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...shivangimorya083
 
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111Sapana Sha
 
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...anilsa9823
 
Production documentary.ppt. x
Production documentary.ppt.               xProduction documentary.ppt.               x
Production documentary.ppt. x21005760
 
Greenery-Palette Pitch Deck by Slidesgo.pptx
Greenery-Palette Pitch Deck by Slidesgo.pptxGreenery-Palette Pitch Deck by Slidesgo.pptx
Greenery-Palette Pitch Deck by Slidesgo.pptxzohiiimughal286
 
What Could Cause Your Subaru's Touch Screen To Stop Working
What Could Cause Your Subaru's Touch Screen To Stop WorkingWhat Could Cause Your Subaru's Touch Screen To Stop Working
What Could Cause Your Subaru's Touch Screen To Stop WorkingBruce Cox Imports
 
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip CallDelhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Callshivangimorya083
 
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girls
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile GirlsVip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girls
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girlsshivangimorya083
 
Chapter-1.3-Four-Basic-Computer-periods.pptx
Chapter-1.3-Four-Basic-Computer-periods.pptxChapter-1.3-Four-Basic-Computer-periods.pptx
Chapter-1.3-Four-Basic-Computer-periods.pptxAnjieVillarba1
 
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Delhi Call girls
 
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdf
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdfSales & Marketing Alignment_ How to Synergize for Success.pptx.pdf
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdfAggregage
 
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Delhi Call girls
 
FULL ENJOY - 9953040155 Call Girls in Sector 61 | Noida
FULL ENJOY - 9953040155 Call Girls in Sector 61 | NoidaFULL ENJOY - 9953040155 Call Girls in Sector 61 | Noida
FULL ENJOY - 9953040155 Call Girls in Sector 61 | NoidaMalviyaNagarCallGirl
 
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Number
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp NumberVip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Number
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Numberkumarajju5765
 
Russian Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...
Russian  Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...Russian  Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...
Russian Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...shivangimorya083
 
How To Troubleshoot Mercedes Blind Spot Assist Inoperative Error
How To Troubleshoot Mercedes Blind Spot Assist Inoperative ErrorHow To Troubleshoot Mercedes Blind Spot Assist Inoperative Error
How To Troubleshoot Mercedes Blind Spot Assist Inoperative ErrorAndres Auto Service
 
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...Hot Call Girls In Sector 58 (Noida)
 
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagardollysharma2066
 
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sex
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sexDelhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sex
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sexshivangimorya083
 

Recently uploaded (20)

Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...
Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...
Hot And Sexy 🥵 Call Girls Delhi Daryaganj {9711199171} Ira Malik High class G...
 
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111
ENJOY Call Girls In Okhla Vihar Delhi Call 9654467111
 
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...
Lucknow 💋 (Genuine) Escort Service Lucknow | Service-oriented sexy call girls...
 
Production documentary.ppt. x
Production documentary.ppt.               xProduction documentary.ppt.               x
Production documentary.ppt. x
 
Greenery-Palette Pitch Deck by Slidesgo.pptx
Greenery-Palette Pitch Deck by Slidesgo.pptxGreenery-Palette Pitch Deck by Slidesgo.pptx
Greenery-Palette Pitch Deck by Slidesgo.pptx
 
What Could Cause Your Subaru's Touch Screen To Stop Working
What Could Cause Your Subaru's Touch Screen To Stop WorkingWhat Could Cause Your Subaru's Touch Screen To Stop Working
What Could Cause Your Subaru's Touch Screen To Stop Working
 
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip CallDelhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
Delhi Call Girls Vikaspuri 9711199171 ☎✔👌✔ Whatsapp Hard And Sexy Vip Call
 
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girls
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile GirlsVip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girls
Vip Hot🥵 Call Girls Delhi Delhi {9711199012} Avni Thakur 🧡😘 High Profile Girls
 
Chapter-1.3-Four-Basic-Computer-periods.pptx
Chapter-1.3-Four-Basic-Computer-periods.pptxChapter-1.3-Four-Basic-Computer-periods.pptx
Chapter-1.3-Four-Basic-Computer-periods.pptx
 
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
 
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdf
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdfSales & Marketing Alignment_ How to Synergize for Success.pptx.pdf
Sales & Marketing Alignment_ How to Synergize for Success.pptx.pdf
 
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
Call Girls in Malviya Nagar Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts Ser...
 
FULL ENJOY - 9953040155 Call Girls in Sector 61 | Noida
FULL ENJOY - 9953040155 Call Girls in Sector 61 | NoidaFULL ENJOY - 9953040155 Call Girls in Sector 61 | Noida
FULL ENJOY - 9953040155 Call Girls in Sector 61 | Noida
 
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Number
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp NumberVip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Number
Vip Hot Call Girls 🫤 Mahipalpur ➡️ 9711199171 ➡️ Delhi 🫦 Whatsapp Number
 
Russian Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...
Russian  Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...Russian  Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...
Russian Call Girls Delhi Indirapuram {9711199171} Aarvi Gupta ✌️Independent ...
 
Call Girls In Greater Noida 📱 9999965857 🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SERVICE
Call Girls In Greater Noida 📱  9999965857  🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SERVICECall Girls In Greater Noida 📱  9999965857  🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SERVICE
Call Girls In Greater Noida 📱 9999965857 🤩 Delhi 🫦 HOT AND SEXY VVIP 🍎 SERVICE
 
How To Troubleshoot Mercedes Blind Spot Assist Inoperative Error
How To Troubleshoot Mercedes Blind Spot Assist Inoperative ErrorHow To Troubleshoot Mercedes Blind Spot Assist Inoperative Error
How To Troubleshoot Mercedes Blind Spot Assist Inoperative Error
 
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...
꧁ ୨⎯Call Girls In Ashok Vihar, New Delhi **✿❀7042364481❀✿**Escorts ServiCes C...
 
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar
83778-77756 ( HER.SELF ) Brings Call Girls In Laxmi Nagar
 
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sex
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sexDelhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sex
Delhi Call Girls Safdarjung 9711199171 ☎✔👌✔Body to body massage with sex
 

Event Data Recorder (EDR) for Self Driving Vehicles

  • 1. Event Data Recorder (EDR) for Self Driving Vehicles Dr. Manish Kochhal, Ph.D. Staff Software Engineer, Autonomous Drive Group NIO Inc opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 2. Outline of the talk What is an EDR? Motivations Implications of Self-Driving Cars on EDR Capturing Data and Decision Making Sensor Fusion – Cameras, Radar, LiDAR, Automatic Control - Drive-by-wire, Brake- by-wire, Steering Assist EDR Implementation – An Embedded and Distributed Real-time System Design Challenge! Standards for EDR format, tools etc. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 3. What is an EDR? • EDR is a tiny blackbox embedded by automakers in the airbag safety systems to record data around the time of the accident. • EDR data is usually kept secret and shared with NHTSA engineers for accident reconstruction, if an investigation occurs. • EDR can also be analyzed for understanding crash dynamics and improving vehicle safety. • EDR blackboxes are not actually mandated. • Privacy concerns have caused the NHTSA to advise automakers only to capture tiny bits of information. Image: autobody-review.com opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 4. EDR – A causal record of events around the time of the accident Image: squarell.com opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 5. EDR Data – In Non-AV Cars Steering Wheel Angle - Steering wheel position, from 0º if wheel is in the neutral position, to +/- 1080º if wheel is turned 3 times to either direction. Lateral Acceleration - A measure of how hard the vehicle is cornering. Lateral acceleration can also be impact induced. Longitudinal Acceleration - A measure of how hard the vehicle is accelerating or decelerating. Longitudinal acceleration can also be impact induced. Yaw Rate - A measure of the rate at which the vehicle is changing its degree of rotation. It is important to note that this measurement is a rate. Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 6. EDR Data - Reports Roll Rate - A measure of the rate at which the vehicle is changing its degree of roll. It is important to note that this measurement is a rate. Delta-V - The change in velocity is calculated from accelerometer data in the Airbag Control Module and presented for longitudinal and lateral directions. In order to determine actual Delta-V for the crash, engineers utilize both datasets to determine the principle direction of force for the crash. Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 7. EDR Data – Countdown to crash Source: Event Data Recorder (EDR) Reports: Understanding the Basics - Expert Article, Robson Forensic, The Expert. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 8. Sensors Publish event data to EDR over CAN Source: Event Data Recorder (EDR) Position Statement, Association for the Advancement of Automotive Medicine Design of ARM based Enhanced Event Data Recorder &Evidence Collecting System, N. P. Sirsikar, P. H. Chandankhede, IOSR-JECE, Volume 9, Issue 5, Ver. V (Sep - Oct. 2014) opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 9. Levels of Vehicle Autonomy Image: Automated Driving Systems, A Vision for Safety, NHTSA opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 10. ASIL - Automotive Safety Integrity Levels • 5-level scale (QM, A, B, C, D) • QM means, “standard Quality Assurance is sufficient” • From ASIL A onwards, additional risk reduction actions must be taken • ASIL D describes the highest risk potential. • ASIL has requirements allocated to it. • The defined safety goals are the top level safety requirements (on vehicle level!) opinions expressed are solely my own and do not express the views or opinions of my employer. Source: https://semiwiki.com/eda/synopsys/7480-functional-safety-is-a-driving-topic-for-iso-26262/
  • 11. NHTSA’s Guideline on Safety for Autonomous Vehicles • NHTSA has some requirements for AEB, LDW, LKA etc. for L2+ autonomous cars. • Following are still evolving: • Federal and State Local Laws • IEEE J3016 has defined what Operational Design Domain (ODD) means – however no accepted standard as to what the ODDs are. • Data Recording standards are defined for L3, L4, and L5. • EDR/Data Recording is important • Assessing car/driver faults under several ODD scenarios • Fail to fallback or takeover by driver or car or both. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 12. EDR of Sensing Data and Decisions Source: Delphi opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 13. EDR needs to capture all decisions opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 14. A screenshot of what the Google car sees approaching a right turn; inset, the view from inside the car. • The baseline image is the detailed area map in grayscale. • Layered atop that are objects identified by the car's sensors, depicted in colorful geometric boxes: purple for vehicles, red for cyclists, yellow for pedestrians. • The red and green ladders are objects that have an immediate impact on the car's speed; in this case, though the traffic light is green, pedestrians prevent a turn, as does a cyclist coming up on the right — in a spot a human driver might easily miss. • The flat green line shows the car's planned route. EDR needs to capture Path Planning, perceived and predicted obstacles, lane markers, construction zones, traffic signs etc. in real-time. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 15. Predicting Bicyclist Intentions • This is an example of a Cruise car trying to predict the intentions of a bicyclist nearby. • Such scenarios involve crunching a lot of sensory inputs for a making real-time conservative decision and accurate predictions. • The data, decisions, and predictions need to be recorded in real-time in the EDR. Complicated prediction of obstacles and their movements need to be recorded. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 16. Advanced BlackBox Should be as advanced as avionics flight data recorders. Should be open, including the data collected, so the truth of self driving limitations is always available to make the next car even safer. Source: https://www.wardsauto.com/industry-voices/level-5-autonomy-it-ai-or-just-deep-programming opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 17. Off-the-shelf Event Data Recorders • 5 Radars • 4 Surround Cameras + 1 Trifocal Camera • 1 LiDAR • 4 Ultrasonic Sensor • 1 IMU • 1 GPS • HD Maps EDR: Capture data from 16+ sensors opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 18. Design Components Involved for EDR Development • Mechanical design of EDR BlackBox • E/E topology • Distributed Communications Middleware – clients publish interesting event data. • Automotive Grade OS and Virtualization – to make applications (including Data Recorder) failsafe, redundant and reliable. • Automotive Grade File System – data integrity vital during read/writes. • Wireless – upload to cloud • Security – EDR cannot be hacked by unauthorized users. • Backup Power – available to save 10-15s of post-crash EDR record • Tools to extract EDR. Automotive OS Features: Containment, Virtualization, Redundancy, Failover opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 19. The Vehicle is a Distributed System opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 20. EDR Framework – System Requirements •Urban city driving – unprotected left turns, pedestrian crossing •Highway driving – lane detection, lane change, •L2/L3+ features: AEB, ACC, LKAS, LDW, Assisted Parking, … •AD Engagement, Disengagement, Decision Uncertainty, Driver Attention, etc … ODD Behavior Specification – specially for L4 and L5 autonomy. •A separate ASIL-D controller •Supports high throughput network connectivity, 1Gbps. •Traffic prioritization – Virtual LAN, flow prioritization based QoS tags •Safety Core in the EDR ECU – to support failover of recording application •High Throughput Read/Writes to in-memory records (Circular Buffers) •High Throughput Read/Writes to ASIL-D filesystem on primary and backup storage arrays. •PCIe or USB – automotive grade driver and stack to interface with memory and storage. •Storage and Memory Requirements to accommodate 16+ sensors recording at varying rates for 10-15 seconds of pre, at, and post accident events. EDR ECU opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 21. ODD Behavior – Classify Events severities: Normal, Medium, High, Critical, Warning, Fault (its Context Sensitive!) ODD Driver behavior Car behavior Environmental behavior Driver engaged Driver Disengaged Driver Assist Driver Manual Driver Asleep AD Engaged AD Disengaged Hands-OnHands-Off Navigation Change Lane Status Pedestrian Objects Detected Potential Obstacles TSR - Status Driver Fallback Object Unclear Safety Decision opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 22. Hypothetical ODD Event Scenario • City Driving Scenario (Most Unpredictable!) • AD Engaged –Normal Event • Driver State, asleep - High • Right turn from start to 2-way street - Normal Event • Traffic Sign Recognition (TSR) event – STOP or Signal – Medium • 4-way stop - High • Decision to take an unprotected Left turn – Critical • Unable to take a left turn – Fault • Able to merge to a lane – Medium • Automatic Emergency Braking – High • Enter Construction Zone/Slow Speed – Medium • Exit Construction Zone/Increase Speed – High • Estimated Unprotected Pedestrian Crossing – Critical • Enter Parking Lot – Medium • Park – Normal • Crash or Safety Stop or AD Disengaged or Driver Takeover – Critical • Airbag Deployed – Fault • Obstacle Classification – Estimate %age – Low, High, Medium, Critical • Decision Certainty – Conservative, otherwise - Low, High, Medium, Critical opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 23. EDR: Event Abstraction Framework Event Clients Interface (Register, Buffer, Trigger, Cancel, Trigger Status) LiDAR Event Client Camera Event Client Navigation Event Client Radar Event Client Ultrasonic Event Client Driver State Event Client Decision Making Event Client Distributed Communications Middleware Object Detection Event Client Event and its Attributes (Priority, Name, Client Id, State) Event Context Airbag Deployment Event Client AEB LDW ACC opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 24. EDR: In-Memory Recording Framework, Circular Buffering to our Rescue! EDR MCU Decision Camera GPS/Maps LiDAR Radar Gateway(CAN/ETH) Pre Post Pre Pre At PostPost Pre CB Organization Primary CBs Secondary CBs Tertiary CBs opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 25. EDR – Blackbox storage on Distributed and Parallel Filesystems • TBs of data depending upon ODD duration and complexity. • We need highly reliable file-system. This is a big data problem and there are existing solutions in data center space. Source: Flash Memory in the emerging age of autonomy, S. Heinrich, Lucid Motors. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 26. Future Work: EDR Format, Extraction and Visualization Tools • Virtual Reality based Accident/Crash Reconstruction based on BlackBox data pulled from the EDR. • Pull the EDR from specific tools provided by the manufacturer. • Convert it to a VR movie – again using the tools provided. Image Source: Courtesy Google Images – Accident Reconstruction opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 27. Existing Questions in Public Domain In a fully autonomous car accident, the EDR data may be the only operational driving record. Can we trust the relationship between automakers and investigators to allow the truth to be revealed? How will we know what the car saw before a crash with a human-driven car or another bicyclist? Currently, there is no independent infrastructure established with guidelines in place to assess the completeness of EDR data. However, there is agreement for L2/L3 like driver assist features. opinions expressed are solely my own and do not express the views or opinions of my employer.
  • 28. Conclusions General acceptable EDR content and format is a WIP Need to get agreement from lawmakers, automotive manufacturers, suppliers, for the EDR content relevant to an ODD. Customers need to be informed about the importance of EDR. EDR Events need to be specified for each of autonomous driving ODD domains. Similar to crash testing, guidelines should be set for applicable ODDs and the EDR recorded. The data recorded is validated. opinions expressed are solely my own and do not express the views or opinions of my employer.