DARPA Grand Challenge

1
DARPA Grand Challenge
It shall be a goal of the Armed Forces that by
2015, one-third of the operational ground combat
vehicles of the Armed Forces are unmanned. (S.
2549, Sec. 217)

• A 175 mile race through the desert

2
(No Transcript)
3
Description of Challenge

• Up to 175 mile route through the desert
• Mountainous terrain up / down 2000 ft
• Completely autonomous vehicles
• For March 2004, 106 teams entered, 25 qualified,
  5 made it past the first 100 yds
• Best showing Carnegie Mellon _at_ 7.4 mi

4
Best Showings

• "Vehicle 22 Red Team (Carnegie Mellon) At mile
  7.4, on switchbacks in a mountainous section,
  vehicle went off course, got caught on a berm and
  rubber on the front wheels caught fire, which was
  quickly extinguished. Vehicle was
  command-disabled.
• "Vehicle 21 SciAutonics II At mile 6.7,
  two-thirds of the way up Daggett Ridge, vehicle
  went into an embankment and became stuck. Vehicle
  was command-disabled.
• "Vehicle 9 The Golem Group At mile 5.2, while
  going up a steep hill, vehicle stopped on the
  road, in gear and with engine running, but
  without enough throttle to climb the hill. After
  trying for 50 minutes, the vehicle was
  command-disabled."

5
Sand
6
Unpaved road
7
Water
8
Rocky Road
9
Rough Terrain
10
Under Highways
11
(No Transcript)
12
Summary of the Rules

• The vehicle must travel autonomously on the
  ground in under ten hours.
• The vehicle must stay within the course
  boundaries as defined by a data file provided by
  DARPA.
• The vehicle may use GPS and other public signals.
  
• No control commands may be sent to the vehicle
  while en route.
• The vehicle must not intentionally touch any
  other competing vehicle. Destructive behavior
  is prohibited.
• Tethered subsystems cannot be propelled or
  maneuvered independently of the ground vehicle.
• Vehicles must have minimal environmental impact.

13
Race Video Footage, 2004

• History Channel Special about the Million Dollar
  Challenge

14
What is our objective?

• Develop technologies for general use
• Invent valuable IP
• Build a new high-value company
• Promote Indiana technology
• Win the 2M prize

15
Current and Future applications of Autonomous
Vehicles

• Delivery and disposal of hazardous materials
• Military logistics -- transport of materials
• Non-military logistics hospitals, airports,
  warehouses, manufacturing plants
• Airport baggage
• Farming and agriculture
• Advance ground warfare
• People movers

16
Major Milestones

• December 15 Basic autonomous operation
• March 11 Application video and paper
• May 2 DARPA Site visit
• September 28 NQE (National Qualifying Event)
• October 8 2005 DARPA Grand Challenge

17
Current Team Members

• Currently 22 All Indiana team members
• 6 entrepreneurial business owners, some with
  substantial robotics experience
• 3 professors with control, robotics, vision
  experience
• 4 students
• 9 engineering/programming professionals
• A combined 300 Years of experience on the team.

18
Where We Meet on Saturdays
19
Skills Needed

• Vendor survey of available sensors and their
  capabilities
• Evaluate trade offs for a "make / buy" decision
  and sensor sophistication level
• Physical mounting, environmental protection and
  articulation of sensor (if required)
• Software signal conditioning and compensation
• Pattern recognition software for terrain,
  obstacles, landmarks and localization
• Sensor integrity analysis software for confidence
  evaluation
• Software interface to navigation controller

20
Sensor Technologies

• DGPS - 3D absolute localization and time standard
  
• LADAR - a "cloud of LASER points" with both
  distance and intensity components
• Vision - Passive cameras stereoscopic, wide
  angle, color
• Odometry - encoders on passive wheels
• Inertial - 6DOF motion sensors
• Magnetic - 3D field orientation, compass
  functions and local fields (power lines)
• RADAR - Background speed, vehicle following /
  passing, metal obstacles
• Thermal - FLIR thermal imaging
• Vibration / Shock - robot health and terrain
  analysis
• Touch - whiskers for "last resort" obstacle
  detection
• Ultrasonic - SONAR for detection of objects
  transparent to other technologies
• Clinometers - gravity based orientation (not
  inertial)
• Helioscope - sun tracking for robot pose and
  shadow identification
• Altimeters - for topographic map rationalization

21
Software

• Motion control software (steering, throttle, etc)
• Real-time image analysis and feature extraction
• Sensor array and fusion software
• Navigation control software
• Path planning
• Obstacle avoidance software
• General system architecture
• Network communications
• Operating system
• Fault tolerance
• Logical deployment of recovery systems
• Mapping technology

22
Mechanical Technologies

• Gyro-stabilized sensor platform
• Electrical system, battery backup
• Environment controls (A/C, humidity)
• Ruggedization
• Robot health
• E-Stop interface
• Vehicle dynamics
• Vehicle protection
• Recovery systems

23
Other Required Functions

• Development tools
• System integration
• Testing
• Project Management
• Cost management
• Marketing
• Public relations
• Web development
• Sponsorship recruitment

24
Jeep Rubicon Chosen, Oct 04
25
(No Transcript)
26
Our Jeep Rubicon
27
Our Jeep Rubicon
28
Installing Drive-by-Wire Now!!
29
Drive-by-Wire
30
Ways for Purdue to be Involved

• Class project(s)
• Extra-curricular activity
• Faculty research
• Other?

31
LADAR Arrays in Hi-res Dynamic Environment
Analysis

• Recognize roads, terrain contours, and identify
  obstacles from an unstable vehicle traveling at
  100 kilometers per hour with resolution in the
  centimeter range

32
Multi-spectral RADAR for Terrain Profiling

• Identify hard and soft terrain features using
  RADAR at different frequencies to identify hard
  packed road paths from surrounding ground and
  rocks from bushes.

33
SONAR arrays in a High Speed, Unstructured
Environment

• Wind and speed create difficulties in extracting
  data from ultrasonic reflections

34
Using Arrays of GPS Receivers to Circumvent
Invalid Positioning

• Differential GPS corrects for a subset of
  positioning errors. An array of GPS receivers
  may allow faulty data from other error sources to
  be identified and either ignored or corrected.

35
Map Matching from Terrain Contours

• Fused sensor data of surrounding terrain is used
  to register against topological and aerial maps
  for road identification and localization without
  GPS.

36
Stabilizing Instrument Platforms on High Speed
Land Vehicles

• Sensor stabilization may be critical for making
  sense of sensor data in real-time.

37
Sensor Fusion for Separating Critical Obstacles
from Benign Terrain Features

• Develop methods for using inputs from multiple
  sensors to separate dangerous obstacles such as
  rocks from easily negotiated terrain features
  such as shadows or tumbleweeds.

38
Sensor Fusion for Identifying Roads in an
Unstructured Environment

• Making best determination of road location with
  oftentimes conflicting imaging and
  obstacle-avoidance sensor data.

39
Strategic Recovery of an Autonomous Vehicle from
Indeterminate Situated State

• If the vehicle is immobile and sensor data is
  contradictory, what steps can be taken to free
  the vehicle from its situation and restore sensor
  effectiveness.

40
Why Purdue Involvement?

• Worthy challenge for top engineering school
• Multi-disciplinary approach
• University-to-university collaboration
• University-to-industry collaboration
• Creation of a high-tech start-up
• Long-term, ground-breaking research opportunities

41
More Information

• Our website is http//IndyRobotics.com
• DARPA site is http//www.darpa.mil/grandchallenge

42
THANK YOU!!

• If you have interest, please contact us!

43
Doug Traster

• Mr. Traster is an electrical engineering graduate
  of the University of Texas at Austin with 30
  years of experience in embedded controller design
  and development, the last ten of which were in
  robotic applications. He is the inventor of three
  patents in video display design two of which are
  licensed for all closed caption televisions. He
  is also knowledgeable in data communications
  technology with patents in HDLC channelized T1.
• Mr. Traster formed his first company, Varix
  Corporation, in 1981. It was one of the first
  personal computer based test equipment companies
  in the world. Varix built a universal PROM and
  PLD programmer, the first of it's kind to contain
  device programming algorithms entirely in
  software.
• Mr. Traster's current company, Volant
  Corporation, was formed in 1985. It operates as a
  turnkey contract developer for a wide range of
  projects including high speed data communication,
  automobile diagnostic equipment, channelized T1,
  database applications, document management, and
  other embedded controller applications. In the
  last 15 years, the company has focused primarily
  on software and electronics for controlling
  robotic equipment - storage / retrieval systems,
  semiconductor device handlers and mobile robots.
• Mr. Traster volunteers his time for Middle School
  and High School robotics. For the last three
  years he has mentored LEGO Mindstorms for the
  Indianapolis Public Library, provided technical
  leadership for several FIRST LEGO League teams
  and provided engineering leadership on a FIRST
  Robotics team.

44
Scott Jones

• Scott Jones, inventor of several patents,
  including telephone-company Voicemail used by
  over 500 million people globally, continues to
  actively build businesses that capitalize on
  advanced technologies. Founder/Chairman of one
  of Indianas leading VC firms, Gazelle
  TechVentures, as well as Founder/Chairman of
  Gracenote (maker of the widely-used Internet
  music recognition service, CDDB) and Escient
  (maker of the first hard-disk based music systems
  for the home), he served as Chairman of the
  Indiana Technology Partnership, a statewide
  organization comprised of business, academic, and
  civic leaders. Graduating from IU in 84, he was
  a research scientist at MITs Artificial
  Intelligence Lab and then founded Boston
  Technology in 86.