Title: Cal Poly Pomona
1Cal Poly Pomona
- Robot Navigation
- Salomón Oldak, Ph.D.
- Electrical and Computer Engineering
- 2/8/06
2DARPA Grand Challenge
- Congress mandate to increase the use of ground
unmanned vehicles. - Congressional Goal 1/3 of armed forces combat
vehicles unmanned by 2015. - Vehicles must be fully autonomous
- Nearly 250 miles on and off road in 10 hours.
3First DARPA Challenge 2004
- First Challenge took place on 3/14/04
- Between Los Angeles and Las Vegas
- 1 Million Prize
- 25 Teams Participated
- CalTech, University of Florida, University of
Alaska, Virginia Tech and others - Palos Verdes High School
4Ghostrider(Blue Team, Berkeley)
5Blue Team Movie
6Obstacles
- Paved Roads
- Overpasses
- Straight-Winding Roads
- Sand, Rock
- Underpasses
- Water
- Natural Obstruction
71st Challenge Results
Failure No Team Completed Task. Max Distance
7.4Mi
82nd DARPA Grand Challenge (10/8/2005)
- Success! 3 Robots Completed task!
9The Winner Stanley
10Stanford University (Sebastian Thrun)
11Sensors Used
- GPS Antenna
- Laser Range Finder (Lidar) (30m)
- Video Camera (80m)
- Odometry (Photo Sensor on Wheel)
12Algorithm
- Problems Vibration would trick sensors to
imagine ghost obstacles, The vehicle thought
its own shadow is an obstacle. - Solution Teach the car. Assess weights to pixels
as a human driver operates the car.
13Future
- 43000 people die in traffic accidents/year in the
US - Robot driven cars will reduce of fatalities
- Accidents can be avoided
- Liability?
14Spirit and Opportunity
15Navigation
- Rovers are mostly teleoperated
- 2 stereo hazard avoidance cameras front and back
- 1 Stereo Navigation camera on mast
- Rovers moves 0.5m at max speed of 34m/h 0.02MPH
16Robot Paradigms
- Paradigm
- A "view" of how things work in the world.
- a set of rules and regulations
- Paradigm Primitives
- SENSE, PLAN, ACT
-
17Paradigms
- Hierarchical
- 1967-1990
- Reactive
- 1988-1992
- Hybrid Deliberative/Reactive
- 1990-
18Hierarchical Paradigm
- The robot operates in a top-down fashion, heavy
on planning. - The robot senses the world, plans the next
action, acts at each step the robot explicitly
plans the next move. - All the sensing data tends to be gathered into
one global world model.
19Shakey (SRI)
First AI Robot (1967-70)
20Assumptions
- Close World World Model Contains everything the
Robot needs to know - Frame Problem The real-world situation is
computationally feasible.
21Problem
- Robots designed under Hierarchical Paradigm were
VERY slow.
22Reactive Paradigm
- Sense-act type of organization.
- The robot has multiple instances of Sense-Act
couplings. - These couplings are concurrent processes, called
behaviours, which take the local sensing data and
compute the best action to take independently of
what the other processes are doing. - The robot will do a combination of behaviours.
23Biological Foundation of Reactive Paradigm
- Reactive Paradigm is based on observations of
ethologists (study of animal behavior) and
cognitive psychology (how humans think and
represent knowledge) - Biology provides existence proofs.
24Behavior Definition (graphical)
BEHAVIOR
Pattern of Motor Actions
Sensory Input
25Arctic Terns
- Arctic terns live in Arctic (black, white, gray
environment, some grass) but adults have a red
spot on beak - When hungry, baby pecks at parents beak, who
regurgitates food for baby to eat - How does it know its parent?
- It doesnt, it just goes for the largest red spot
in its field of view (e.g., ethology grad student
with construction paper) - Only red thing should be an adult tern
- Closer large red
26Types of Behaviors
- Reflexive
- stimulus-response, often abbreviated S-R (like
knee tapped). Hardwired - Reactive
- learned or muscle memory. (Riding a bike,
skiing, etc.) - Conscious
- deliberately stringing together (Building a
Robot)
WARNING Overloaded terms Roboticists often use
reactive behavior to mean purely reflexive, And
refer to reactive behaviors as skills
27Example Cockroach Hide
- light goes on, the cockroach turns and runs
- when it gets to a wall, it follows it
- when it finds a hiding place (thigmotrophic),
goes in and faces outward - waits until not scared, then comes out
- even if the lights are turned back off earlier
28Behaviors are Concurrent
29What happens when theres a conflict from
concurrent behaviors?
- Equilbrium
- Feeding squirrels-feed, flee hesitate in-between
- Dominance
- Sleepy, hungry either sleep or eat
- Cancellation
- Sticklebacks defend, attack build a nest
?
30Reactive Robots
RELEASER
behavior
SENSE
ACT
- Most apps are programmed with this paradigm
- Biologically based
- Behaviors (independent processes), released by
perceptual or internal events (state) - No world models or long term memory
- Highly modular, generic
- Overall behavior emerges
31Example My Real Baby
- Behaviors?
- Touch-gt Awake
- Upside down Awake-gt Cry
- Awake Hungry -gt Cry
- Awake Lonely -gt Cry
- Note can get crying from multiple behaviors
- Note internal state (countdown timer on Lonely)
32Subsumption ArchitectureRodney Brooks
From http//www.spe.sony.com/classics/fastcheap/in
dex.html
33 Runaway
34Example Perception Polar Plot
if sensing is ego-centric, can often eliminate
need for memory, representation
- Plot is ego-centric
- Plot is distributed (available to whatever wants
to use it) - Although it is a representation in the sense of
being a data structure, there is no memory
(contains latest information) and no reasoning
(2-3 means a wall)
35Potential Fields (Example Navigation)
7
6
5
4
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0
0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2
36Wavefront Algorithm
7
6
5
4
3
2
1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
18 17 16 15 14 13 12 11 10 9 9 9 9 9 9 9
17 17 16 15 14 13 12 11 10 9 8 8 8 8 8 8
17 16 16 15 14 13 12 11 10 9 8 7 7 7 7 7
17 16 15 15 1 1 1 1 1 1 1 1 6 6 6 6
17 16 15 14 1 1 1 1 1 1 1 1 5 5 5 5
17 16 15 14 13 12 11 10 9 8 7 6 5 4 4 4
17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 3
17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2
(0,7) -gt (1,7) -gt (2,7) -gt (3,7) -gt (4,7) -gt
(5,7) -gt (6,7) -gt (7,7) -gt (8,7) -gt (9,7) -gt
(10,7) -gt (10,6) -gt (11,6) -gt (11,5) -gt (12,5) -gt
(12,4) -gt (12,3) -gt (13,3) -gt (13,2) -gt (14,2) -gt
(14,1) -gt (15,1) -gt (15,0)
37Hybrid Deliberate/Reactive Paradigm
- The robot first plans (deliberates) how to best
decompose a task into subtasks (also called
mission planning) and then what are the
suitable behaviours to accomplish each subtask. - Then the behaviours starts executing as per the
Reactive Paradigm. - Sensing organization is also a mixture of
Hierarchical and Reactive styles sensor data
gets routed to each behaviour the needs that
sensor, but is also available to the planner for
construction of a task-oriented global world
model.
38Deliberation v.s. Planning
- Besides planning robot has to perform other
tasks such as map making, performance
monitoring, learning, etc. - All these tasks together with planning are known
as DELIBERATION
39Sensing Organization
Deliberative functions Can eavesdrop Can have
their own Sensors Have output which Looks like
a sensor Output to a behavior (virtual sensor)
40Hybrid Behaviors
- Behaviors are extended to
- Reflexive
- Innate
- Learned
- (Just like in ethology)
41Architectures Common Functionality
- Mission planner
- Cartographer
- Sequencer agent
- Behavioral manager
- Performance monitor/problem solving agent (fairly
rare)
42Several Hybrid Approaches Have Been Developed
- AuRA (Arkin 1986)
- Atlantis (Gat 1991)
- Sensor-Fusion Effects (SFX) (Murphy 1996)
- 3-Tiered (3T) (JPL1990s)
- Saphira (Konolige 1998)
- Tack Control Architecture (Simmons 1997)
- Planner-Reactor (Lyons and Hendriks 1992)
- Procedural Reasoning System (PRS) (Georgeff and
Lansky 1987) - SSS (Connell 1992)
- Multi-Valued Logic (Saffiotti 1995)
- SOMASS Hybrid Assembly System (Malcom and
Smithers 1990) - Agent Architecture (Hayes-Roth 1993)
- Etc., Etc.
43Georgia Tech TMR(Tactical Mobile Robot) Robots
44NAVIGATION
- Topological Navigation Qualitative Navigation
- Metric Navigation Quantitative Navigation
- Navigation Algorithm usually is part of
Deliberative part of Hybrid Architecture.
45Qualitative Navigation uses Landmarks
46floor plan
Gateway is an opportunity to change path heading
relational graph
Relational Methods Nodes landmarks,
gateways, goal locations Edges navigable path
47Quantitative Navigation
- Want to get from one point to another with an
optimization criteria - Minimize Time
- Minimize Energy
- Minimize Distance
- Etc.
48Space Representation Voronoi Graphs
- Imagine a fire starting at the boundaries,
creating a line where they intersect,
intersections of lines are nodes - Result is a relational graph
49Space Representation Rectangular Graph
50A Algorithm
51Discussion
Questions ?
52References
- Murphy R.R. Introduction to AI Robotics, MIT
Press, 2000 - http//www.policyalmanac.org/games/aStarTutorial.h
tm (Accessed 2/6/06) - http//robots.stanford.edu/ (Accessed (2/6/06)
- http//www.ghostriderrobot.com/index.php?idrobot
(Accessed 2/6/06)