Navigating the potential disruption of driverless cars in Australia - - PowerPoint PPT Presentation

▶
navigating the potential disruption of driverless cars in
SMART_READER_LITE
LIVE PREVIEW

Navigating the potential disruption of driverless cars in Australia - - PowerPoint PPT Presentation

Jun 15, 2023 160 likes •621 views

Navigating the potential disruption of driverless cars in Australia Professor Michael Milford ww www.robo boti ticv cvisio sion. n.org ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION roboticvision.org Disclaimer The views and opinions

slide-1
SLIDE 1

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Navigating the potential disruption of driverless cars in Australia

Professor Michael Milford

slide-2
SLIDE 2

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Disclaimer

  • The views and opinions expressed in

this presentation are those of the author and do not necessarily reflect the official policy or position of any other agency, employer or organization.

  • All information presented is general in

nature and does not take into account your personal or organization’s situation.

  • All information is provided without

guarantee on the part of the presenter.

  • The presenter disclaims any liability in

connection with the use of this information.

slide-3
SLIDE 3

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

What’s Up (2019 update) What Can Go Wrong Australian Activity Current Public and Industry Attitudes

slide-4
SLIDE 4

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Self-driving Cars

slide-5
SLIDE 5

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Why?

Self-Driving Cars

slide-6
SLIDE 6

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Why?

2.4 die per minut 2.4 die per minute

slide-7
SLIDE 7

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

History

NavLab 1, CMU (1986) https://www.youtube.com/watch?v=ntIczNQKfjQ NavLab 5, CMU (1996) https://www.youtube.com/watch?v=bdQ5rsVgPuk DARPA Grand Challenge, 2004‐2005 https://www.youtube.com/watch?v=M2AcMnfzpN DARPA Urban Grand Challenge, 2007 https://www.youtube.com/watch?v=M2AcMnfzpN

slide-8
SLIDE 8

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Current State of Play

slide-9
SLIDE 9

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Autonomy Levels

http tps: s://w //www.the heverge.com com/2016/9 /9/2 /28/1 8/13076948 6948/sel /self- f-dr driving- g-car car-po poll-au auton

  • nomy-kel

elley-bl blue-book

  • k

Where the t Where the techn chnology logy is curre is currently

slide-10
SLIDE 10

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Current Challenges and Incidents

slide-11
SLIDE 11

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Lives Saved and Taken as Deployment Increases

https://www.youtube.com/watch?v=wsI7N7hhZTg

slide-12
SLIDE 12

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

May 7th, 2016: Tesla Model S Truck Crash

“Neithe either A Autopilo pilot t nor r the dri the driver n er notice ticed d the whit the white side of side of the tract the tractor r traile trailer r against a against a brig brightly lit sky ly lit sky, so the so the brak brake e was not applied” as not applied” http tps: s://electrek //electrek.co/20 co/2016/0 6/07/0 /01/u 1/under derstan standing- ing-fatal- tal-tesla-acciden sla-accident-au t-autop

  • pilo

ilot-nhtsa- t-nhtsa-probe/ be/

slide-13
SLIDE 13

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Visual Sensing Challenge

slide-14
SLIDE 14

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

March 18, 2018: Uber Fatality

slide-15
SLIDE 15

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Uber Fatality: What Happened?

  • Radar and LIDAR detect
  • bject 6 seconds before

impact

  • Classification order:

unknown object – vehicle – bicycle.

  • 1.3 seconds before impact:

emergency braking communicated to car, not supervisor driver

  • Uber had disabled Volvo

AEB

  • Supervising human driver

responds < 1 second before impact

https://www.theverge.com/2018/5/24/17388696/uber‐self‐driving‐crash‐ntsb‐report

slide-16
SLIDE 16

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

March 23, 2018: Tesla Model X fatal crash

Source: NTSB/S. Engleman https://abc7news.com/automotive/i‐team‐exclusive‐victim‐who‐ died‐in‐tesla‐crash‐had‐complained‐about‐auto‐pilot/3275600/

  • 3 seconds prior, Autopilot

speeds up from 62 to 70.8 mph (set to 75 mph)

  • No evasive maneuverers
  • No driver hands on wheel for

previous 6 seconds

slide-17
SLIDE 17

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

The Big Challenge: Human Interaction

https://www.youtube.com/watch?v=VG68SKoG7vE

slide-18
SLIDE 18

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Conventional Startup Example: NuTonomy

https:/ https://www www.y .youtube utube.com/

  • m/watc

atch?v= ?v=iP_ iP_lAjIf lAjIfZwU wU

  • Par

Partner nerships ships for v r vehicles, ri hicles, ride de sharing mark sharing market, go governm rnment ent

  • Contr

Controlle lled d ride sharing flee ride sharing fleet

  • Formal rigor

rmal rigorous us appr approa

  • ach
  • People

eople do don’t n’t ow

  • wn c

n cars

  • Singa

Singapor pore, e, Bos Boston…

slide-19
SLIDE 19

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Conventional Startup Example: Cruise Automation

  • Rapid

Rapidly sca y scaling ing

  • So

Software and re and testing-int sting-intensiv ensive e appr approa

  • ach
slide-20
SLIDE 20

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

ZooX

slide-21
SLIDE 21

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Car Maker Corporate Example: Toyota & Toyota Research Institute

ht https://www tps://www.y .youtu

  • utube.com/w

be.com/watch? atch?v=5s6HbrO =5s6HbrOYads ds

  • Long

Long term rm outlo

  • utlook,

k, cult cultura ural co context

  • Vehi

hicle m cle making legacy cy

  • Du

Dual al assi assist stiv ive ( e (Guardian) and ian) and full fully aut autonomous pr proje

  • jects

(Cha (Chauffeur) ur)

  • Pri

Privat ate c e car r owner

  • wnership sti

ship still ll fe feasible

slide-22
SLIDE 22

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Tech Company Corporate Example: Waymo/Google

https:/ https://www www.y .youtube utube.com/

  • m/watc

atch?v= ?v=uHb uHbMt6W Mt6WDhQ8 DhQ8

  • Long

Long his histor

  • ry
  • New builder

er o

  • f ve

vehicl cles es

  • Exis

Existing ng t tech ch and cons and consum umer infra infrastructure 20 2012 20 2018

slide-23
SLIDE 23

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Google Waymo (2019 Updates)

slide-24
SLIDE 24

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Between Industry and Academia Example: Vedecom

slide-25
SLIDE 25

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Between Industry and Academia Example: Vedecom

slide-26
SLIDE 26

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Public and Industry Attitudes

slide-27
SLIDE 27

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Public Expectation

slide-28
SLIDE 28

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

The Moral Machine

http://moralmachine.mit.edu/

slide-29
SLIDE 29

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

General Backpedalling and Expectation Management by the Industry

slide-30
SLIDE 30

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Diversification, Pivots and Specialization

slide-31
SLIDE 31

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Niche Area Play Example: Nuro.AI

  • Automated delivery
  • Partnering with major supermarket chains
  • Point of differentiation: several exploitable advantages with no

passengers

slide-32
SLIDE 32

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Niche Area Play Example: Nuro.AI

https://www.youtube.com/watch?v=XKXbacNQGI8

slide-33
SLIDE 33

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Niche Area Play Example: Perceptive Automata

https://www.youtube.com/watch?v=D‐2gmeTUijc

slide-34
SLIDE 34

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Guidelines for Assesing, Reacting and Planning

slide-35
SLIDE 35

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Though Experiment: Investigate what Phones are Capable of in 2019

  • 3 megapixel single rear

camera

  • Camera doesn’t work in

low light

  • Lasts 15 days on standby
  • 2.4 inch 76,800 pixel

display

  • 42 MB of storage
  • Dual ARM9 220 MHz

https://www.mobilegazette.com/2006‐review‐06x12x22.htm

slide-36
SLIDE 36

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Though Experiment: Investigate what Phones are Capable of in 2019

  • 3 rear cameras: 16MP +

12MP + 12MP, normal, ultrawide, telephoto

  • Camera as good as human

eye in low light

  • Lasts a couple of days on

standby

  • 6.4 inch 4,400,000 pixel

display

  • 512 GB of storage
  • Eight core Exynos: 2x2.7

GHz, 2x2.3 GHz and 4x1.9 GHz

  • Incredible range of powerful

applications

https://www.techradar.com/au/ reviews/samsung‐galaxy‐s10‐ plus

  • 3 megapixel single

rear camera

  • Camera doesn’t work

in low light

  • Lasts 15 days on

standby

  • 2.4 inch 76,800 pixel

display

  • 42 MB of storage
  • Dual ARM9 220 MHz
slide-37
SLIDE 37

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

What you Can and Can’t Learn with Unmodified, Off the Shelf, Commercially Available Technology

  • Get a snapshot of what currently

commercially available tech can do, right now.

  • Gain organizational experience
  • Learn general information about

coding, sensing, compute

  • Learn general insights about the

interaction of a complex system with the world

  • Give you any insights into what actual

current proprietary technology can or can’t do

  • Give you insights into the difference

between momentary challenges and fundamental challenges

  • Predict what could happen in a few

years time

  • Learn the nuts and bolts of how

modern AI and sensing

slide-38
SLIDE 38

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

How Automated Vehicles Will Interact With Road Infrastructure Now and in the Future

slide-39
SLIDE 39

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

slide-40
SLIDE 40

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Improvement of “out of the box” technologies

  • 1. “Out of the box”
  • 2. Intermediate System
  • 3. Final System

Modern learning‐based systems can learn where to “expect” lines

slide-41
SLIDE 41

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Key Takeaways

slide-42
SLIDE 42

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

My Recommendations About Self-driving Cars & Related Technologies

Avoid the temptation to think in unchanging absolutes about the technology e.g. “never be like a human”, “solves everything” See through the hype but don’t dismiss everything out of hand Keep an informed, regularly updated awareness of the range of scenarios that could play out

slide-43
SLIDE 43

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

We’re Always Here to Talk and Collaborate!

slide-44
SLIDE 44

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Demonstrating self‐driving cars and robots at QUT Robotronica Talking robotics on Channel Ten’s The Project PhD Student James demoing self‐driving cars on BrainBuzz Self‐driving cars with Stan Grant on ABC’s Matter of Fact

We’re Always Here to Talk and Collaborate!

slide-45
SLIDE 45

ww www.robo boti ticv cvisio sion. n.org

roboticvision.org

ARC CENTRE OF EXCELLENCE FOR ROBOTIC VISION

Navigating the potential disruption of driverless cars in Australia

Professor Michael Milford