Transcript The Hierarchical Paradigm - City University of New York

7
Chapter 7:
Hybrid Deliberative/Reactive
Paradigm
Part 1: Overview & Managerial Architectures
Part 2: State Hierarchy Architectures
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
1
7
Objectives
• Describe the hybrid paradigm in terms of 1) SPA and 2) sensing
organization
• Given a list of responsibilities, be able to say whether it belongs in
the deliberative layer or in the reactive layer
• List the five basic components of a Hybrid architecture: sequencer
agent, resource manager, cartographer, mission planner,
performance monitoring and problem solving agent.
• Be able to describe the difference between managerial, state
hierarchy, and model-oriented styles of Hybrid architectures.
• Be able to describe the use of state to define behaviors and
deliberative responsibilities in state hierarchy styles of Hybrid
architectures
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
2
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
What can a reactive architecture do? What can’t it do?
Path planning, handling detours due to blockage, map making,
learn from past rescues
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
3
7
Organization: Plan, Sense-Act
PLAN
SENSE
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
ACT
Chapter 7: Hybrid Deliberative/Reactive Paradigm
4
7
Sensing Organization
SENSOR 3
SENSOR 1
Deliberative functions
*Can “eavesdrop”
*Can have their own
WORLD MAP/
Sensors
KNOWLEDGE REP
*Have output which
Looks like a sensor
virtual sensor Output to a behavior
(virtual sensor)
BEHAVIOR
BEHAVIOR
SENSOR 2
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
ACTUATORS
BEHAVIOR
Chapter 7: Hybrid Deliberative/Reactive Paradigm
5
7
Deliberation v. Reaction
as a function of TIME
• Past, Present, Future
• Reactive
– exists in the PRESENT (will a bit of
duration)
• Deliberative
– can reason about the PAST
– can project into the FUTURE
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
6
7
Architectures: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?
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
7
7
Architectures:
Common Functionality
• Mission planner
• Cartographer
• Sequencer agent
• Behavioral manager
• Performance monitor/problem solving
agent (fairly rare)
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
8
7
Architectures: 3 Styles
• Managerial (division of responsibilities looks
like in business)
– AuRA, SFX
• State Hierarchies (strictly by time scope)
– 3T
• Model-Oriented (models serve as virtual
sensors)
– Saphira, TCA
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
9
7
Mgr Architecture 1:
AuRA (Autonomous Robot Arch.)
Ron Arkin, Georgia Institute of Technology
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
10
7
AuRA Architectural Layout
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
11
Architectures:
Common Functionality
7
•
•
•
•
•
Mission planner
Cartographer
Sequencer agent
Behavioral manager
Performance monitor/problem solving
agent
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
12
7
AuRA Architectural LayoutPerformance
Cartographer
Monitoring
Mission
Planner
Sequencer
Behavioral
manager
(mgr+schemas)
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Emergent behavior
Chapter 7: Hybrid Deliberative/Reactive Paradigm
13
7
HOW WOULD THIS DO USAR TASK?
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
14
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
What can a reactive architecture do? What can’t it do?
Path planning, handling detours due to blockage, map making,
learn from past rescues
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
15
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
16
7
Example USAR (overlay)
• Cartographer accepts the map
• Navigator uses path planning
algorithm to visit nodes in order
of likelihood of survivors
• Pilot determines the list of
behaviors, Motor Schema
Manager instantiates them (MS
& PS) and waits for termination
• Homeostatic might notice that
robot is running out of power,
so opportunistically picks up
low probability room on way
back to home
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
17
7
Mgr. Architecture 2:
SFX (Sensor Fusion Effects)
• Focus on sensing
• Biomimetic organization
• deliberative layer consists of managerial agents
• reactive layer has tactical behaviors
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
18
7
SFX (Sensor Fusion Effects)
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
19
7
SFX (Sensor Fusion Effects)
Recognition
perception
Cartographer
(model/map
making)
Deliberative
Layer Managers
Cerebral
Cortex-like
functions
Choice of behaviors, resource
allocation, motivation, context
Sensor
Behavioral
Whiteboard
Whiteboard
Parameters to behaviors,
sensor failures, task progress
Behaviors
Sense
Sense
Sense
Receptive
actions
Field
Sense
Sense
Sense
Sense
Sensor
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Superior
Colliculus-like
functions
(using direct
perception, fusion)
Focus of attention,
recalibration
Muscle
Muscle
Muscle
Actuators
Chapter 7: Hybrid Deliberative/Reactive Paradigm
20
7
SFX Implementation
HOW WOULD THIS DO USAR?
Cartographer
C++
Interface
Sensors
Sensors
Sensors
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Effector Mgr
Information
directs sensing
Sensor Mgr
Mgr
Behaviors
Behaviors
Behaviors
Acuators
Acuators
Acuators
Chapter 7: Hybrid Deliberative/Reactive Paradigm
Behaviors
Use, Fuse
Lisp
Task Planner
21
7
Ability to Substitute Components
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
22
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
23
7
Example USAR (overlay)
• Cartographer accepts the map
• Task Planner agent asks for
path, requests behaviors,
passes to managerial layer
• Sensing and Effector Mgrs
negotiate allocation
• Behaviors run until terminate
or encounter exception (either
preset condition by mgrs or
through monitoring)
• Mgrs can see “below” but not
above--cannot relax constraint
of Planner/Boss
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
24
7
Tactical Behaviors
sensors
strategic behaviors
inclinometer
camera
tactical behaviors
actuators
slope
safe velocity
follow-path
strategic
velocity
clutter
drive
motor
speed-control
direction to path
safe direction
steer
motor
avoid
sonar
obstacles
how much vehicle turns
swivel camera
center-camera
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
camera
pan
motor
Chapter 7: Hybrid Deliberative/Reactive Paradigm
25
7
UGV Competition 1997
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
26
7
Summary:
Managerial Architectures
• How does the architecture distinguish between reaction and
deliberation?
– Deliberation: global knowledge or world models, projection
forward or backward in time
– Reaction: behaviors which have some past/persistence of
perception and external state
• How does it organize responsibilities in the deliberative portion?
– hierarchy of managerial responsibility, managers may be peer
software agents
• How does overall behavior emerge?
– From interactions of a set of behaviors dynamically
instantiated and modified by the deliberative layer
– assemblages of behaviors
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
27
7
Chapter 7:
Hybrid Deliberative/Reactive
Paradigm
Part 1: Overview & Managerial Architectures
Part 2: State Hierarchy & Model-Based Architectures
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
28
7
Objectives
• Describe the hybrid paradigm in terms of 1) SPA and 2)
sensing organization
• Given a list of responsibilities, be able to say whether it
belongs in the deliberative layer or in the reactive layer
• List the five basic components of a Hybrid architecture:
sequencer agent, resource manager, cartographer, mission
planner, performance monitoring and problem solving agent.
• Be able to describe the difference between managerial, state
hierarchy, and model-oriented styles of Hybrid architectures.
• Be able to describe the use of state to define behaviors and
deliberative responsibilities in state hierarchy styles of Hybrid
architectures
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
29
7
Plan, Sense-Act
PLAN
SENSE
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
ACT
Chapter 7: Hybrid Deliberative/Reactive Paradigm
30
7
Sensing Organization
SENSOR 3
WORLD MAP/
KNOWLEDGE REP
virtual sensor
SENSOR 1
BEHAVIOR
BEHAVIOR
SENSOR 2
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
ACTUATORS
BEHAVIOR
Chapter 7: Hybrid Deliberative/Reactive Paradigm
31
7
Architectures: 3 Styles
• Managerial (division of responsibilities looks like in
business)
– AuRA, SFX
• State Hierarchies (strictly by time scope or “state”)
– 3T
• Model-Oriented (models serve as virtual sensors)
– Saphira, TCA
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
32
7
State Hierarchy Architectures
• How does the architecture distinguish between reaction and
deliberation?
– Deliberation: requires PAST or FUTURE knowledge
– Reaction: behaviors are purely reflexive and have only local,
behavior specific; require only PRESENT
• How does it organize responsibilities in the deliberative portion?
– By internal temporal state
• PRESENT (controller)
• PAST (sequencer)
• FUTURE (planner)
– By speed of execution
• How does overall behavior emerge?
– From generation and monitoring of a sequence of behaviors
– assemblages of behaviors called skills
– subsumption
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
33
7
3T Architecture
• Used extensively at NASA
• Merging of subsumption
variation (Gat, Bonasso),
RAPs (Firby), and vision
(Kortenkamp)
• Has 3 layers
Dave Kortenkamp,
TRAC Labs (NASA JSC)
– reactive
– deliberative
– in-between (reactive planning)
• Arranges by time
• Arranges by execution rate
– ex. vision in deliberation
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
34
7
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
35
Mission planner
7
Performance
monitor
cartographer
sequencer
Behavior
mgr
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Emergent
Chapter 7: Hybrid Deliberative/Reactive
Paradigm
behavior
36
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
What can a reactive architecture do? What can’t it do?
Path planning, handling detours due to blockage, map making,
learn from past rescues
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
37
7
Model-Oriented Architectures
• How does the architecture distinguish between reaction and
deliberation?
– Deliberation: anything relating a behavior to a goal or objective
– Reaction: behaviors are “small control units” operating in
present, but may use global knowledge as if it were a sensor
(virtual sensor)
• How does it organize responsibilities in the deliberative portion?
– Behavioral component
– Model of the world and state of the robot
– throwback to Hierarchical Paradigm with global world model
but virtual sensors
– Deliberative functions
• How does overall behavior emerge?
– From generation and monitoring of a sequence of behaviors
– voting or fuzzy logic for combination
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
38
7 Saphira Architecture
• Developed at SRI by
Konolige, Myers, Saffioti
• Comes with Pioneer
robots
• Behaviors produce fuzzy
outputs, fuzzy logic
combines them
• Has a global rep called a
Local Perceptual
Structure to filter noise
• Instead of RAPs, uses
PRS-Lite
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
39
7
Saphira and LPS
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
40
7
Sequencer agent,
Mission Planning,
Performance mon.
Behavior mgr
Cartographer
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Emergent behavior
Chapter 7: Hybrid Deliberative/Reactive Paradigm
41
7
Symbol-Grounding Problem
• Computers (and AI) reasons using
symbols
– Ex. “room”, “box,” “corner,” “door”
• Robots perceive raw data
• How to convert sensor readings to these
labels?
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
42
7
Spatial World Knowledge
• What do you see?
• How could a robot
reliably extract the
same labels?
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
43
7
Types of Knowledge (Arkin)
•
•
•
•
•
•
Spatial World knowledge
Object knowledge
Perceptual knowledge
Behavioral knowledge
Ego knowledge
Intentional knowledge
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
44
7
Task Control Architecture
• Developed by Reid Simmons
• Used extensively by CMU
Field Robotics Projects
– NASA’s Nomad, Ambler,
Dante
• Closest to Hierarchical in
philosophy, but strong
reactive theme showing up in
implementation
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
45
7
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
TCA
Chapter 7: Hybrid Deliberative/Reactive Paradigm
46
7
Motivating Example for
Deliberation: USAR
• Worker places robot at entrance
to unstable building, loads in the
floor plan, contextual knowledge
and tells robot to look for
survivors efficiently and map out
safe routes for workers to pass
through
• contextual knowledge includes
probability of where people are
more likely to be
What can a reactive architecture do? What can’t it do?
Path planning, handling detours due to blockage, map making,
learn from past rescues
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
47
7
Evaluation of Hybrids
• Support of Modularity: high
• Niche targetability: high (ex. Lower levels of
AuRA, SFX, 2 1/2 T is just reactive)
• Robustness: SFX and 3T explicitly monitor
• Think in closed world, act in open world
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
48
7
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
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
49
7
Class Exercise
• Form groups of 3
• Design Hostage Rescue robot software
– What are the key tasks? Robot capabilities?
Environment?
– Do you need to know more?
– What paradigm? What architectural style?
• Is there deliberation or just reaction?
• Which paradigm?
Introduction to AI Robotics (MIT Press),
copyright Robin Murphy 2000
Chapter 7: Hybrid Deliberative/Reactive Paradigm
50