ELEC-E8111 Autonomous Mobile Robots CONTROL …
Transcript of ELEC-E8111 Autonomous Mobile Robots CONTROL …
![Page 1: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/1.jpg)
ELEC-E8111 – Autonomous Mobile Robots
CONTROL ARCHITECTURES
Mika Vainio, Research Fellow, Aalto
![Page 2: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/2.jpg)
Contents of the lecture
• Motivation and state-of-the-art issues
• Short introduction to control paradigms
• The Hierarchical Paradigm
• Biological foundations for the reactive paradigm
• The Reactive Paradigm
• The Hybrid Paradigm
2
![Page 3: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/3.jpg)
3
MAINLY BASED ON THESE TWO BOOKS BY RONALD ARKIN AND ROBIN MURPHY
![Page 4: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/4.jpg)
Robotics (along with information technology and
biotechnology) is identified as one of
the key technologies of the
future (the new millennium)
CEO Honda
![Page 5: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/5.jpg)
What is a robot?
• A robot is a re-programmable, multi-functional, manipulator designed tomove material, parts, or specialized devices though variable programmedmotions for the performance of a task (Robotics Industry Association)
• An intelligent robot is a machine able to extract information from itsenvironment and use knowledge about its world to move safely in ameaningful and purposeful manner. (Arkin)
• A robot is a system which exists in the physical world and autonomouslysenses its environment and acts in it. (Mataric)
![Page 6: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/6.jpg)
What makes a MOBILE robot?
sensors
effectors/actuators
locomotion system
on-board computer system
6
![Page 7: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/7.jpg)
Why is Robotics so hard in reality?
Sensors are limited and crude
Effectors are limited and crude
State (internal and external, butmostly external) is partially-observable
Environment is dynamic (changingover time)
Environment is full of potentially-useful information
7
![Page 8: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/8.jpg)
Key Issues
• Grounding in reality: not just planning in an
abstract world
• Situatedness: tight connection with the
environment
• Embodiment: having a body
• Emergent behavior: interaction with the
environment
• Scalability: increasing task and environment
complexity
8
![Page 9: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/9.jpg)
9
REAL WORLD TESTING - SpotMini by Boston Dynamics
![Page 10: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/10.jpg)
What is autonomy?
• the ability to make one's own decisions
and act on them
• for robots, the ability to sense the
situation and act on it appropriately
10
![Page 11: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/11.jpg)
…Autonomy
• Autonomy can be complete, as inautonomous robots, or partial, as in tele-operated robots.
• examples of autonomous robots: R2D2
• examples of tele-operated robots: NASA'srobots before Pathfinder
• sliding autonomy
11
![Page 12: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/12.jpg)
From scifi...
FILM by S. Kubrick
2001: A Space Odyssey
(1968)
FILM by J. Cameron
The Terminator (1985)
REALLY SHORT HISTORY
![Page 13: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/13.jpg)
...to factories
ABB Industrial manipulator robot
In 1961 the first industrial robot, Unimate, joined the assembly line at a General Motors plant.
![Page 14: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/14.jpg)
Automatically Guided Vehicles
![Page 15: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/15.jpg)
From volcanoes down to abyss
![Page 16: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/16.jpg)
To other planets...
![Page 17: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/17.jpg)
and even to our homes.
![Page 18: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/18.jpg)
THE (fantastic) FUTURE: 4 focal areas
![Page 19: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/19.jpg)
Military robotics. Scary stuff. Skipping this one.
![Page 20: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/20.jpg)
20
Japanese birth and death rates
Finnish population pyramid in 2017
FOCUS 1
AGING
POPULATIONS
![Page 21: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/21.jpg)
In hospital, elderly and disabled care
Social aspects,
logistics, surgery,
rehabilition,etc.
Left: Paro Therapeutic Robot by AIST
Middle: HelpMate robotic materials transport system by HelpMate Robotics, Inc
Right: da Vinci Surgical System by Intuitive Surgical
![Page 22: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/22.jpg)
SELF-DRIVING CARS
FOCUS 2
![Page 23: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/23.jpg)
Started 1984
The Robotics
Institute at the
School of Computer
Science, Carnegie
Mellon University
In July 1995, the
team took Navlab5
from Pittsburgh to
San Diego on a
proof-of-concept trip,
dubbed "No Hands
Across America",
with the system
navigating for all but
50 of the 2850 miles,
averaging over 60
MPH.
FOUNDATION – NAVLAB project
![Page 24: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/24.jpg)
24
The first competition of the DARPA Grand Challenge was held on
March 13, 2004 in the Mojave Desert region of the United States,
along a 150-mile (240 km) route. None of the robot vehicles finished
the route. No winner was declared, and the cash prize was not
given.
The second competition of the DARPA Grand Challenge began at 6:40am
on October 8, 2005. All but one of the 23 finalists in the 2005 race
surpassed the 11.78 km (7.32 mi) distance completed by the best vehicle
in the 2004 race. Five vehicles successfully completed the 212 km
(132 mi) course.
QUANTUM
LEAP-
DARPA Grand
Challenge
![Page 25: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/25.jpg)
![Page 26: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/26.jpg)
![Page 28: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/28.jpg)
![Page 29: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/29.jpg)
![Page 30: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/30.jpg)
ENTENSIVE TESTING
![Page 31: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/31.jpg)
Supportive
infra:
Self-healing
maps
(HERE)
![Page 32: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/32.jpg)
![Page 33: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/33.jpg)
SENSIBLE 4MADE IN FINLAND
![Page 34: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/34.jpg)
![Page 35: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/35.jpg)
35
![Page 36: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/36.jpg)
Drones –
the last
mile
problem
36
FOCUS 3
Amazon Prime
![Page 37: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/37.jpg)
37
By Roksenhorn - Own work, CC BY-SA 4.0
Zipline
![Page 38: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/38.jpg)
Humanoids
ATLAS by Boston
Dynamics
FOCUS 4
![Page 39: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/39.jpg)
CONTROL ARCHITECTURES
THE REAL DEAL
39
![Page 40: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/40.jpg)
Control
Robot control refers to the way in which
the sensing and action of a robot are
coordinated. The many different ways in
which robots can be controlled all fall
along a well-defined spectrum of control.
40
![Page 41: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/41.jpg)
DELIBERATIVE REACTIVE
Purely Symbolic Reflexive
SPEED OF RESPONSE
PREDICTIVE CAPABILITIES
DEPENDENCE ON ACCURATE, COMPLETE WORLD MODELS
Representation-dependent
Slower response
High-level intelligence (cognitive)
Variable latency
Representation-free
Real-time response
Low-level intelligence
Simple computation
41
![Page 42: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/42.jpg)
42
Architecture
provides a principled way of organizing a control system. However, in addition to providing structure, it imposes constraints on the way the control problem can be solved [Mataric]
describes a set of architectural components and how they interact [Dean & Wellman]
![Page 43: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/43.jpg)
43
Evaluating an Architecture
• support for modularity: does it show good
software engineering principles?
• niche targetability: how well does it work for
the intended application?
• ease of portability to other domains: how
well would it work for other applications or
other robots?
• robustness: where is the system vulnerable,
and how does it try to reduce that
vulnerability?
![Page 44: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/44.jpg)
Control Approaches:
• Reactive Control : Don’t think, (re)act.
• Deliberative Control : Think hard, act later.
• Hybrid Control : Think and act
independently, in parallel.
44
![Page 45: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/45.jpg)
Control Trade-offs:
• Thinking is slow.
• Reaction must be fast.
• Thinking enables looking head (planning) to avoid
bad solutions.
• Thinking too long can be dangerous (e.g., falling off
a cliff, being run over).
• To think, the robot needs (a lot of) accurate
information => world models.
45
![Page 46: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/46.jpg)
Reactive Systems:
• Don’t think, react!
• Reactive control is a technique for tightlycoupling perception (sensing) andaction, to produce timely roboticresponse in dynamic and unstructuredworlds.
• Think of it as "stimulus-response".
• A powerful method: many animals arelargely reactive.
46
![Page 47: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/47.jpg)
Limitations:
• Minimal (if any) state.
• No memory.
• No learning.
• No internal models /
representations of the world.
47
![Page 48: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/48.jpg)
Deliberative Systems
• Based on the sense->plan->act model
• Inherently sequential
• Planning requires search, which is slow
• Search requires a world model
• World models become outdated
• Search and planning takes too long
48
![Page 49: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/49.jpg)
Everybody’s got plans…
until they get hit
-Mike Tyson
49
![Page 50: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/50.jpg)
Hybrid Systems
• Combine the two extremes
– reactive system on the bottom
– deliberative system on the top
– connected by some intermediate layer
• Often called 3-layer systems
• Layers must operate concurrently
• Different representations and time-scales
between the layers
• The best of both worlds?
50
![Page 51: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/51.jpg)
51
The Hierarchical (aka deliberative)
Paradigm
![Page 52: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/52.jpg)
52
Hierarchical Paradigm…
• Top-down: plan, plan, plan
• Control-theoretic: must measure error in order to control device
• Planning means:dependence on world models
![Page 53: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/53.jpg)
53
Juha Backman currently @ Natural Resources Institute Finland (LUKE)
2013
![Page 54: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/54.jpg)
54
Organization
World model:
1. A priori rep
2. Sensed info
3. Cognitive
PLANSENSE ACT
![Page 55: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/55.jpg)
55
Nested Hierarchical Controller
(Meystel)
![Page 56: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/56.jpg)
56
NHC Planner
![Page 57: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/57.jpg)
57
SPA = Planner-based
SPA has serious drawbacks
• Problem 1: Time-Scale
• Problem 2: Space
• Problem 3: Information
• Problem 4: Use of Plans
![Page 58: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/58.jpg)
58
Problem 1: Time-Scale
• It takes a very (prohibitively) long time to
search in a real robot's state space, as that
space is typically very large
• Real robots may have collections of simple
digital sensors (e.g., switches, IRs), a few
more complex ones (e.g., cameras), or
analog sensors (e.g., encoders, gauges,
etc.) => "too much information”
=> Generating a plan is slow.
![Page 59: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/59.jpg)
59
Problem 2: Space
• It takes a lot of space (memory) to represent
and manipulate the robot's state space
representation.
• The representation must contain all
information needed for planning. =>
Generating a plan can be large.
• Space is not nearly as much of a problem as
time, in practice.
![Page 60: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/60.jpg)
60
Problem 3: Information
• The planner assumes that the
representation of the state space is
accurate and up-to-date =>
The representation must be constantly
updated and checked
• The more information, the better =>
"too little information"
![Page 61: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/61.jpg)
61
Problem 4: Use of Plans
The resulting plan is only useful if:
a) the environment does not change during
the execution of a plan in a way that
affects the plan
b) the representation was accurate enough
to generate a correct plan
c) the robot's effectors are accurate enough
to perfectly execute each step of the plan
in order to make the next step possible.
![Page 62: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/62.jpg)
62
Planners Live On in Robotics
• The SPA approach has not been abandoned,
it has been expanded
• Given the two fundamental problems with
purely deliberative approaches, we can
augment them:
– search/planning is slow, so save/cache important
and/or urgent decisions;
– be ready to respond or re-plan when the plan fails.
![Page 63: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/63.jpg)
63
Biological Foundations of the Reactive
Paradigm
![Page 64: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/64.jpg)
64
Behavior Definition
BEHAVIOR
B
S: Sensory
Input
R: Pattern
of Motor
Actions
(responses)
A behavior is a mapping of sensory inputs to a
pattern of motor actions which are then used to
achieve a task.
Notation: B(S)=R
![Page 65: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/65.jpg)
65
Types of Behaviors
• Reflexive
– stimulus-response, often abbreviated S-R
• Reactive
– learned or “muscle memory”
• Conscious
– deliberately stringing together
WARNING Overloaded terms:
Roboticists often use “reactive behavior” to mean purely reflexive,
And refer to reactive behaviors as “skills”
![Page 66: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/66.jpg)
66
Ethology: Coordination and Control of
Behaviors
Nobel 1973 in
-von Frisch
-Lorenz
-Tinbergen
www.nobel.se
INNATE RELEASING MECHANISMS
![Page 67: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/67.jpg)
67
Motivating Example:
Arctic Terns
• Arctic terns live in Arctic (black, white, gray
environment, some grass) but adults have a red
beak
• When hungry, baby pecks at parent’s beak, who
regurgitates food for baby to eat
• How does it know its parent?
– It doesn’t, 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
![Page 68: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/68.jpg)
68
Behavior template
BEHAVIOR
Sensory
Input
Pattern
of Motor
Actions
![Page 69: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/69.jpg)
69
“the feeding behavior”
Feeding
BEHAVIOR
Sensory
Input
Pattern
of Motor
Actions
S=[RED] R= PECK AT RED
B(S)=R
![Page 70: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/70.jpg)
70
“the feeding releaser”
Feeding
BEHAVIOR
RED PECK AT RED
Releaser
internal state
RED &
HUNGRYsensory input
][)],([,, feedingBREDATPECKRHUNGRYREDS
![Page 71: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/71.jpg)
71
Innate Releasing Mechanisms
BEHAVIOR
Sensory
Input
Pattern
of Motor
Actions
Releaser
Sensory input
and/or
internal state
![Page 72: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/72.jpg)
72
Example: 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, then comes out
• even if the lights are turned back off earlier
![Page 73: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/73.jpg)
73
Observation 1:
Fixed Pattern Action
• light goes on, the cockroach turns and runs
• when it gets to a wall, it follows it
• when it finds a hiding place, goes in and faces outward
• waits, then comes out
• even if the lights are turned back off earlier
![Page 74: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/74.jpg)
74
Observation 2: Exhibits Taxis
• 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
to light
to wall
to
niche
![Page 75: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/75.jpg)
75
Break into Behaviors
• light goes on, the cockroach turns and runs
• when it gets to a wall, it follows it on right
• when it finds a hiding place, goes in and faces outward
• waits until not scared, then comes out
Flee
Follow-
wall
hide
flee
wallfollow
hide
![Page 76: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/76.jpg)
76
Cockroach Hiding: the behaviors
hide
wallfollow
flee
B
![Page 77: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/77.jpg)
77
Find Releasers
• light goes on, the cockroach
turns and runs
• when it gets to a wall, it follows
it on right
• when it finds a hiding place,
goes in and faces outward
• waits until not scared, then
comes out
Flee
Follow-
wall
hide
LIGHT
present?N
Y
SCARED &
SURROUNDED
present?N
BLOCKED &
SCARED
present?N Ooops, need internal state:
Scared
![Page 78: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/78.jpg)
78
Internal State Set
• light goes on, the cockroach is scared and turns and runs
• when it gets to a wall, it follows it on right
• when it finds a hiding place, goes in and faces outward
• waits until not scared, then comes out
Flee
Follow-
wall
hide
LIGHT
present?N
Y
SCARED &
SURROUNDED
present?N
BLOCKED &
SCARED
present?N
SCARED
![Page 79: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/79.jpg)
79
Cockroach Hiding
hide
wallfollow
flee
B
)&(
)&),(
SCAREDSURROUNDED
SCAREDRIGHTONBLOCKED
LIGHT
S angle
![Page 80: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/80.jpg)
80
Action (Responses)• light goes on, the
cockroach gets scared, turns and runs
• when it gets to a wall, it follows it
• when it finds a hiding place, goes in and faces outward
• waits until not scared,then comes out
Flee
Follow-
wall
hide
LIGHT
present?N
Y
SCARED &
SURROUNDED
present?N
BLOCKED &
SCARED
present?N
SCARED
steer random,
drive forward
steer =F(deg to wall)
drive forward const.
steer =F(deg to wall)
drive forward const.
stop
![Page 81: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/81.jpg)
81
Cockroach Hiding
hide
wallfollow
flee
B
)&(
)&),(
SCAREDSURROUNDED
SCAREDRIGHTONBLOCKED
LIGHT
S angle
)),(%(
))(),(%(
))(,(
slowdrivesurroundedfsteerr
SCAREDfdriveblockedfsteerr
SCAREDfdriverandomsteerr
R
hide
wallfollow
flee
r
r
r
R
![Page 82: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/82.jpg)
82
What happens when there’s a
conflict from concurrent behaviors?
? )*( RGC
![Page 83: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/83.jpg)
83
• If the rules are not triggered by mutually-exclusive
conditions, more than one rule can be triggered in
parallel, resulting in two or more different actions
being output by the system.
• Deciding among multiple actions or behaviors is
called arbitration, and is in general a difficult
problem.
You need arbitration
![Page 84: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/84.jpg)
84
Arbitration can be done based on:
• a fixed priority hierarchy (processes have pre-
assigned priorities)
• a dynamic hierarchy (process priorities change
at run-time)
• learning (process priorities may be initialized
or not, and are learned at run-time, once or
repeatedly/dynamically)
![Page 85: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/85.jpg)
85
Schema Theory
SCHEMA is used in cognitive science and ethology to refer to a
particular organized way of perceiving cognitively and responding
to a complex situation or set of stimuli
• a behavior is a schema, consists of
– perceptual schema
– motor schema
– other behaviors
![Page 86: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/86.jpg)
86
Fly Snapping
![Page 87: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/87.jpg)
Ex. Fly Snapping Behavior IRM
snap(blob)track(blob)
p=x,y,z,
=100%
snap,
100%small moving
dark blob
Releaser:
small moving dark blob
![Page 88: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/88.jpg)
88
Schema Instantiation (SI)
snap(blob)track(blob)
x,y,z,
100%
snap,
100%
Releaser:
small moving dark blob
present? N
Y
/dev/null
![Page 89: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/89.jpg)
89
Schema/Schema Instantiation
behavior
schema
releaser
present?N
1.
snap(blob)track(blob)track(blob) snap(blob)
snap(blob)track(blob)
Perceptual
Schema Library
Motor
Schema Library
2.
snap(blob)track(blob)
x,y,z,
100%
3.
![Page 90: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/90.jpg)
90
Advantages
• modular
• can assemble new behaviors from existing schemas– learning by experimentation
• can substitute alternatives– reroute nerves
![Page 91: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/91.jpg)
91
Back to Toads and Frogs: Instantiation
for each eye
snap,
100%
snap(blob)track(blob)
x,y,z,
100%
Releaser:
small moving dark blob
present?N
Y
/dev/null
snap,
100%
snap(blob)track(blob)
x,y,z,
100%
Releaser:
small moving dark blob
present?N
Y
/dev/null
Left eye
Right eye
Snap at
vector sum
(middle)
![Page 92: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/92.jpg)
92
General Principles
• All animals possess a set of behaviors
• Releasers for these behaviors rely on both internal
state and external stimulus
• Perception is filtered; perceive what is relevant to the
task
![Page 93: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/93.jpg)
93
The Reactive Paradigm
![Page 94: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/94.jpg)
94
Review: Lessons from Biology
• Programs should decompose complex actions into behaviors. Complexity emerges from concurrent behaviors acting independently
• Agents should rely on straightforward activation mechanisms such as IRM
• Perception filters sensing and considers only what is relevant to the task (action-oriented perception)
• Behaviors are independent but the output may be used in many ways including: combined with others to produce a resultant output or to inhibit others
![Page 95: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/95.jpg)
95
Hierarchical Organization is
“Horizontal”
![Page 96: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/96.jpg)
96
More Biological is “Vertical”
![Page 97: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/97.jpg)
97
Reactive Robots
• 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
SENSE ACT
RELEASERbehavior
![Page 98: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/98.jpg)
98
Example 1: Robomow
• Behaviors?
• Random
• Avoid
– Avoid(bump=obstacle)
– Avoid(wire=boundary)
• Stop
– Stop(tilt=ON)
• All activewww.friendlymachines.com
![Page 99: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/99.jpg)
Steps in Designing a Reactive
Behavioral System
99
Describe the task
Describe the robot
Describe the environment
Specification
& Analysis:
ecological
niche
Implement & refine each behavior
Test each behavior independently
Implementation
& unit testing
Test behaviors together System
Testing
Describe how the robot should
act in response to its environmentDesign
![Page 100: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/100.jpg)
Introduction to
AI Robotics
(MIT Press)
Chapter 5: Designing a Reactive Implementation 100
Design Tool: Behavior TableAn agent is attracted to light. If it sees light, it heads in that direction.
If it encounters an obstacle, it turns either left or right, favoring the
direction of the light. If there is no light, it sits and waits. But if an
obstacle appears, the agent runs away.
1. photropism
2. Obstacle avoidance
![Page 101: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/101.jpg)
Introduction to
AI Robotics
(MIT Press)
Chapter 5: Designing a Reactive Implementation 101
Behavior Table
Releaser Behavior Motor
Schema
Percept Perceptual
Schema
Light phototropism move2Light()
:Attraction
Light:
direction &
strength
Brightest(di
r), atLight()
Range
<tasked
level>
Obstacle
avoidance
avoid(): turn
left or right;
runaway()
proximity Obstacle()
An agent is attracted to light. If it sees light, it heads in that direction.
If it encounters an obstacle, it turns either left or right, favoring the
direction of the light. If there is no light, it sits and waits. But if an
obstacle appears, the agent runs away.
![Page 102: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/102.jpg)
102
Reactive: 2 main styles
• Historically, there are two main styles of creating a
reactive system
– Subsumption architecture
• Layers of behavioral competence
• How to control relationships
– Potential fields
• Concurrent behaviors
• How to navigate
• They are equivalent in power
![Page 103: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/103.jpg)
103
Subsumption:
Rodney Brooks
https://people.csail.mit.edu/brooks/
Panasonic
Professor of
Robotics
(emeritus) at MIT.
![Page 104: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/104.jpg)
104
The Subsumption Architecture (Brooks 1985)
![Page 105: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/105.jpg)
105
The Subsumption Architecture
• systems are built from the bottom up
• components are task-achieving actions/behaviors (not functional
modules)
• components can be executed in parallel
• components are organized in layers, from the bottom up
• lowest layers handle most basic tasks
• newly added components and layers exploit the existing ones
• each component provides and does not disrupt a tight coupling
between sensing and action
• there is no need for internal models: "the world is its own best
model"
![Page 106: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/106.jpg)
106
• Subsumption systems grow from the bottom
up, and layers can keep being added,
depending on the tasks of the robot.
• How exactly layers are split up depends on
the specifics of the robot, the environment,
and the task.
• There is no strict recipe, but some solutions
are better than others, and most are derived
empirically.
![Page 107: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/107.jpg)
107
• The inspiration behind the Subsumption
Architecture is the evolutionary process, which
introduces new competencies based on the
existing ones.
• Complete creatures are not thrown out and new
ones created from scratch; instead, solid, useful
substrates are used to build up to more complex
capabilities.
![Page 108: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/108.jpg)
108
• The original Subsumption Architecture was
implemented using a language based on finite
state machines (FSMs) augmented with a very
small amount of state (AFSMs), themselves
implemented in Lisp.
• An AFSM can be in one state at a time, can
receive one or more inputs, and send one or more
outputs AFSMs are connected by communication
wires, which pass input and output messages
between them.
![Page 109: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/109.jpg)
Situated Automata
• A formal notion of finite state machines whose
inputs are connected to sensors and whose
outputs are connected to effectors are called
situated automata.
• Situated means existing in and interacting with a
complex world, and automata is the formal name
for FSMs (formally: finite state automata).
• Situated automata are used to create reactive
principled control systems.
109
![Page 110: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/110.jpg)
FSA Diagrams
• States and state transitions are most easily
encoded in finite state automata and drawn as
finite state diagrams
• The states of the diagrams can also be
behaviors, so the diagrams show sequences
of behavior transitions
• These are called finite state acceptors
• Acceptor M is a quadruple (Q,d,q0,F)
110
![Page 111: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/111.jpg)
Finite State Acceptors
• M (Q, d,q0,F):
– Q is the set of legal behavioral states
– d is a transition function from a state and an
input to the next state (can be represented as
a table)
– q0 is the starting behavioral configuration
– F is the set of accepting states (a subset of
Q, indicates completion)
111
![Page 112: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/112.jpg)
Introduction to
AI Robotics
(MIT Press)
Chapter 5: Designing a Reactive Implementation 112
FSA: M={Q,d,q0,F}
• Q: all the states, each is “q”- behaviors
• d: transition function, d(q,s)= new behavior
• q0: Start state(s)- part of Q
• F: Terminating state(s)- part of Q
![Page 113: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/113.jpg)
113
Arbitration in Subsumption
• Arbitration: deciding who has control
• Inhibition: prevents output signals from reaching
effectors
• Suppression: replaces input signal with the
suppressing message
• The above two are the only mechanisms for
coordination => Results in priority-based
arbitration, the rule or layer with higher priority takes
over, i.e., has control of the AFSM
![Page 114: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/114.jpg)
114
Behavioral ModuleInput
Wires
Inhibitor
Suppressor
Output
Wires
Reset
I
S
R
![Page 115: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/115.jpg)
115
Designing in Subsumption
• Qualitatively specify the overall behavior
needed for the task
• Decompose that into specific and independent
behaviors (layers)
• The layers should be bottom-up and
consisting of disjoint actions
• Ground low-level behaviors in the robot’s
sensors and effectors
• Incrementally build, test, and add
![Page 116: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/116.jpg)
116
Subsumption Philosophy
• Modules should be grouped into layers of competence
• Modules in a higher lever can override or subsume behaviors in the next lower level
– Suppression: substitute input going to a module
– Inhibit: turn off output from a module
• No internal state in the sense of a local, persistent representation similar to a world model.
![Page 117: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/117.jpg)
117
Subsumption Evaluation
• Strengths:
– Reactivity (speed, real-time nature)
– Parallelism
– Incremental design => robustness
– Generality
• Weaknesses:
– Inflexibility at run-time
– Expertise needed in design
![Page 118: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/118.jpg)
Potential Fields:
Ronald Arkin
https://www.cc.gatech.edu/aimosaic/faculty/arkin/index.html
Regents' Professor, Director
of Mobile Robot Laboratory
School of Interactive
Computing, College of
Computing, Georgia Tech
![Page 119: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/119.jpg)
119
Potential Fields Philosophy
• The motor schema component of a behavior can be expressed with a potential fields methodology– A potential field can be a “primitive” or constructed from
primitives which are summed together– The output of behaviors are combined using vector summation
• From each behavior, the robot “feels” a vector or force– Magnitude = force, strength of stimulus, or velocity– Direction
• But we visualize the “force” as a field, where every point in space represents the vector that it would feel if it were at that point
![Page 120: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/120.jpg)
120
Schema Theory (as presented earlier)
SCHEMA is used in cognitive science and ethology to refer to a
particular organized way of perceiving cognitively and responding
to a complex situation or set of stimuli
• a behavior is a schema, consists of
– perceptual schema
– motor schema
– other behaviors
![Page 121: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/121.jpg)
121
Motor Schemas
• Motor schemas are a type of behavior encoding:
They are based on schema theory (Arbib);
Provide large grain modularity; Distributed
concurrent schemas used; Based on
neuroscience and cognitive sci.
• Represented as vector fields
• Composed into “assemblages” by fusion (not
competition)
![Page 122: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/122.jpg)
122
ES1
ES2
ES3
INTERNAL
SENSORS
MOTORS
TRANSMITTER
SCHEMAS
VECTOR
ROBOTMOTOR SCHEMASENVIRONMENTAL
SENSORS
EN
VIR
ON
ME
NT
BROADCAST
MEDIUM
RS1 IS1
IS2
TS1
TS2
RS2 RS3
PS1
PS3
PS2
Key:
RS - Receptor Schema
TS - Transmitter Schema
PS - Perceptual Schema
MS - Motor Schema
IS - Internal Sensor
ES - Environmental Sensor
![Page 123: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/123.jpg)
123
Schema Representation
• Responses represented in uniform vector
format
• Combination through cooperative coordination
via vector summation
• No predefined schema hierarchy
• Arbitration is not used - gain values control
behavioral strengths
![Page 124: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/124.jpg)
124
Designing with Schemas
• Characterize motor behaviors needed
• Decompose to most primitive level, use biological
guidelines where appropriate
• Develop formulas to express reaction
• Conduct simple simulations
• Determine perceptual needs to satisfy motor schema
inputs
• Design specific perceptual algorithms
• Integrate/test/evaluate/iterate
![Page 125: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/125.jpg)
125
Example: Run Away via Repulsion
![Page 126: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/126.jpg)
126
5 Primitive Potential Fields
a) Uniform b) Perpendicular c) Attractive d) Repulsive e) Tangential
![Page 127: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/127.jpg)
127
Combining Fields for
Emergent Behavior
obstacleobstacle
goal
If robot were dropped anywhere on this grid,
it would want to move to goal and avoid obstacle:
Behavior 1: MOVE2GOAL
Behavior 2: RUNAWAY
The output of each independent behavior is a vector,
the 2 vectors is summed to produce emergent behavior
obstacle
goal
![Page 128: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/128.jpg)
128
Note: In this example, repulsive field only extends for 2 meters;
the robot runs away only if obstacle within
2 meters
Note: in this example, robot can sense the
goal from 10 meters away
Fields and Their Combination
![Page 129: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/129.jpg)
129
Path Taken
• If robot started at this location, it would take the following path
• It would only “feel”the vector for the location, then move accordingly, “feel” the next vector, move, etc.
• Pfield visualization allows us to see the vectors at all points, but robot never computes the “field of vectors” just the local vector
Robot only feels
vectors for this
point when it (if)
reaches that point
![Page 130: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/130.jpg)
130
Example: follow-corridor or follow-
sidewalk
Perpendicular Uniform
Combined
Note use of
Magnitude profiles:
Perpendicular decreases
![Page 131: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/131.jpg)
131
But how does the robot see a wall
without reasoning or intermediate
representations?
• Perceptual schema “connects the dots”, returns relative
orientation
PS:
Find-wall
MS: Perp.
MS: UniformS
Sonars
orientation
![Page 132: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/132.jpg)
132
Strengths and Weaknesses
• Advantages– Easy to visualize– Easy to build up software libraries
– Fields can be parameterized
– Combination mechanism is fixed, tweaked with gains– Support for modularity and parallelism
– Run-time flexibility
• Disadvantages– Local minima problem (sum to magnitude=0)
• Box canyon problem
– Jerky motion
![Page 133: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/133.jpg)
HYBRID SOLUTIONS
133
![Page 134: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/134.jpg)
Inventing Hybrid Control
• The basic idea is simple: we want the
best of both worlds (if possible)
• That means to combine reactive and
deliberative control
• This implies combining the different
time-scales and representations
• This mix is called hybrid control
134
![Page 135: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/135.jpg)
Key Questions
• How does the architecture distinguish between reaction and deliberation?
• How does it organize responsibilities in the deliberative portion?
• How does overall behavior emerge?
![Page 136: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/136.jpg)
Common Functionalities
• Mission planner
• Cartographer
• Sequencer agent
• Behavioral manager
• Performance monitor/problem solving agent (fairly rare)
![Page 137: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/137.jpg)
Organization: Plan, Sense-Act
SENSE
PLAN
ACT
![Page 138: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/138.jpg)
Introduction to
AI Robotics
(MIT Press),
copyright
Robin Murphy
2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm 138
Sensing Organization
BEHAVIOR
BEHAVIOR
BEHAVIOR
SENSOR 1
SENSOR 2
ACTUATORS
WORLD MAP/
KNOWLEDGE REPSENSOR 3
virtual sensor
Deliberative functions
*Can “eavesdrop”
*Can have their own
Sensors
*Have output which
Looks like a sensor
Output to a behavior
(virtual sensor)
![Page 139: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/139.jpg)
Organizing Hybrid Systems
• A hybrid system typically consists of three
components:
– a reactive layer
– a planner
– a layer that puts the two together
• Hybrid architectures are often called three-
layer architectures
• The planner and the reactive system are
both standard, as we have covered them
so far
139
![Page 140: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/140.jpg)
The Magic Middle
• The middle layer has a hard job:
– 1) compensate for the limitations of both the
planner and the reactive system
– 2) reconcile their different time-scales
– 3) deal with their different representations
– 4) reconcile any contradictory commands between
the two
• This is the challenge of hybrid systems =>
achieving the right compromise between the
two ends
140
![Page 141: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/141.jpg)
Dynamic Re-planning
• Reaction can influence planning
• Any "important" changes discovered by the
low-level controller are passed back to the
planner in a way that the planner can use to
re-plan
• The planner is interrupted when even a partial
answer is needed in real-time
• The reactive controller (and thus the robot) is
stopped if it must wait for the planner to tell it
where to go.
141
![Page 142: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/142.jpg)
Planner-Driven Reaction
• Planning can influence reaction
• Any "important" optimizations the planner
discovers are passed down to the reactive
controller
• The planner’s suggestions are used if they are
possible and safe => Who has
priority, planner or reactor?
• It depends, as we will see...
142
![Page 143: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/143.jpg)
Universal Plans
• Suppose for a given problem, all possible
plans are generated for all possible situations
in advance, and stored (e.g., automated tic-
tack-toe)
• If for each situation a robot has a pre-existing
optimal plan, it can react optimally, be
reactive and optimal, it has a universal plan
(These are complete reactive mappings)
143
![Page 144: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/144.jpg)
Viability of Universal Plans
• A system with a universal plan is reactive;
the planning is done at compile-time, not at
run-time
• Universal plans are not viable in most
domains, because:
– the world must be deterministic
– the world must not change
– the goals must not change
– the world is too complex (state space is too large)
144
![Page 145: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/145.jpg)
Introduction to
AI Robotics
(MIT Press),
copyright
Robin Murphy
2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm 145
Hybrid Summary
• P,S-A, deliberation uses global world models, reactive uses behavior-specific or virtual sensors
• Architectures generally have modules for mission planner, sequencer, behavioral mgr, cartographer, and performance monitoring
• Deliberative component is often divided into sub-layers (sequencer/mission planner or managers/mission planner)
• Reactive component tends to use assemblages of behaviors
![Page 146: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/146.jpg)
Relative Strengths (Hybrid Control)
• Deliberative planners– Rely heavily on world models
– Can readily integrate world knowledge
– Have broader perspective and scope
• Reactive & behavior-based systems– Afford modular development
– Provide real-time robust performance in dynamic world
– Provide for incremental growth
– Tightly coupled to incoming sensory data
146
![Page 147: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/147.jpg)
Example: AuRA
• Roland C. Arkin (1986)
• Planning is viewed as configuration
• Initial A* planner integrated with schema-
based controller
• Provides modularity, flexibility, and adaptability
147
![Page 148: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/148.jpg)
AuRA Architectural Layout
![Page 149: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/149.jpg)
AuRA Architectural Layout
Cartographer
Sequencer
Mission
Planner
Behavioral
manager
(mgr+schemas)
Performance
Monitoring
Emergent behavior
![Page 150: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/150.jpg)
ADDITIONAL JOURNAL PAPER
150
Ronald C. Arkin & Tucker Balch (1997) AuRA: principles and
practice in review, Journal of Experimental & Theoretical
Artificial Intelligence, 9:2-3, 175-189, DOI:
10.1080/095281397147068
![Page 151: ELEC-E8111 Autonomous Mobile Robots CONTROL …](https://reader030.fdocuments.us/reader030/viewer/2022041000/624eeeb6c7d9425fae3eb1fa/html5/thumbnails/151.jpg)
React when you can, plan when you must.
151
-Tom Mitchell, CMU
Robot control refers to the way in which the sensing and action of a robot
are coordinated. The many different ways how robots can be controlled all
fall along a well-defined spectrum of control ranging from purely symbolic,
detailed representation dependent planning based systems all the way to
reactive, representation-free real-time solutions.
Robot architecture provides a structured and principled way of organizing a
control system. It defines the components, their interactions, prevailing
constraints and operations how to survive and achieve the given mission
objectives.