The Hierarchical Paradigm - PowerPoint PPT Presentation

1 / 36
About This Presentation
Title:

The Hierarchical Paradigm

Description:

The Hierarchical Paradigm – PowerPoint PPT presentation

Number of Views:34
Avg rating:3.0/5.0
Slides: 37
Provided by: robinm53
Category:

less

Transcript and Presenter's Notes

Title: The Hierarchical Paradigm


1
The Hierarchical Paradigm
  • Describe the Hierarchical Paradigm in terms of
    the 3 robot primitives and its organization of
    sensing
  • Name and evaluate one representative Hierarchical
    architecture in terms of support for modularity,
    niche targetability, ease of portability to other
    domains, robustness
  • Solve a simple navigation problem using STRIPS
    (hint work through Sec. 2.2.2)
  • Understand precondition, closed world assumption,
    open world, frame problem
  • List two advantages and disadvantages of the
    Hierarchical Paradigm

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
2
Organization
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
3
Shakey
  • First AI robot
  • Built by SRI (Stanford Research Institute) for
    DARPA 1967-9
  • Used Strips as main algorithm for controlling
    what to do

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
4
Strips Means-ends analysis
Go to Stanford AI Lab
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
  INITIAL STATE Tampa, Florida (0,0)
GOAL STATE Stanford, California (1000,200)
Difference 1020 miles
5
Difference Table
 
dlt200 miles
FLY
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
100ltdlt200
TRAIN
dlt100
DRIVE
modedifference_table(INITIAL STATE, GOAL STATE,
difference)
  • Look up what to do FLY
  • Not at SAIL, so repeat
  • Look up what to do DRIVE

6
Preconditions
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
dlt200 miles
FLY
100ltdlt200
TRAIN
dlt100
DRIVE (rental)
 
DRIVE (personal car)
 
How do I know if Im at the airport or at
home? Now must keep up with the state of the world
7
Maintaining State of the WorldAdd and Delete
Lists
 
 
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
dlt200 miles
FLY
 
100ltdlt200
TRAIN
 
dlt100
DRIVE (rental)
at airport
 
 
 
DRIVE (personal)
at home
 
 
 
 
8
Class Exercise
 
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
  • Write down the world model, the operator applied,
    the change in world state, etc. to go from Tampa
    to Stanford

 
9
Strips Summary
  • Designer must set up
  • World model representation
  • Difference table with operators, preconditions,
    add delete lists
  • Difference evaluator
  • Strips assumes closed world
  • Closed world world model contains everything
    needed for robot (implication is that it doesnt
    change)
  • Open world world is dynamic and world model may
    not be complete
  • Strips suffers from frame problem
  • Frame problem representation grows too large to
    reasonably operate over

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
10
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
  • Types of architectures Levis, George Mason
    University
  • operational architecture describes what the
    systems does, not how it does it
  • systems architecture describes how a system
    works in terms on major subsystems
  • technical architecture implementation details

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
11
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?

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
12
Hierarchical Paradigm
  • Top-down
  • Plan, plan, plan
  • Control-theoretic
  • must measure error in order to control device
  • Planning means
  • dependence on world models

13
Nested Hierarchical Controller(Meystel)
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
14
NHC Planner
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
15
RCS (Albus)
  • the hierarchy
  • how the hierarchy works for navigation
  • how it is implemented
  • nodes and modules
  • planning time periods

16
Examples of RCS Apps
Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
17
RCS-4 Levels
18
Each Level has a RCS Node
Engineering of Mind, Albus Mystel, 2001
Sensory Processing, World Modeling, Behavior
Generation, Value Judgment
19
Implementation ViewNodes are Recursive
20
Demo III XUV
http//museum.nist.gov/exhibits/timeline/item.cfm?
itemId38
Experimental Unmanned Vehicle in action at Ft.
Indiantown Gap. Photo courtesy of the Army
Research Labs. Nov. 2001
21
Demo III Control Hierarchy
PLANNER vehicle1 vehicle2
Section 10m
VEHICLE PLANNER communications plan AM plan
RSTA plan
Vehicle 1m
Subsystem 5s
AM PLANNER Driver Plan Gaze
plan
COMMS PLANNER message list
RSTA PLANNER gaze plan
DRIVER PLANNER Velocity Plan
GAZE PLANNER Stereo Gaze Plan LADAR Gaze Plan
Primitive 500ms
VELOCITY PLANNER F Wheels R Wheels F Steer R Steer
Servo 50ms
F Wheel
R Wheel
F Steer
F Steer
22
RCS XUV Example
Vehicle Level AM Plan (A1A10) Primitive
Level Driver Plan (D1D10)
23
t0.5
Primitive Level Driver Plan extends to A2
24
t1 s
Obstacle Detected Primitive Level Driver
Plan new waypoints Vehicle Level detects too
large a variation
25
t1 s Vehicle Level Planner Opt 1
Vehicle Level new AM Plan Primitive Level new
Driver Plan
26
t3
More obstacle is seen fail upwards
again Vehicle Level new AM Plan Primitive
Level new Driver Plan
27
t4.5s
Vehicle Level new AM Plan skip A1, go to A2
Primitive Level new Driver Plan
28
t6 s
Vehicle Level new AM Plan skip old A2
Primitive Level new Driver Plan
29
Exercise Adapt to Rescue Robots?
PLANNER vehicle1 vehicle2
Section 10m
VEHICLE PLANNER communications plan AM plan
RSTA plan
Vehicle 1m
Subsystem 5s
AM PLANNER Driver Plan Gaze
plan
COMMS PLANNER message list
RSTA PLANNER gaze plan
DRIVER PLANNER Velocity Plan
GAZE PLANNER Stereo Gaze Plan LADAR Gaze Plan
Primitive 500ms
VELOCITY PLANNER F Wheels R Wheels F Steer R Steer
Servo 50ms
F Wheel
R Wheel
F Steer
F Steer
30
Nodes are made from Modules
G is feedback
xd is from above
actions
sensors
transform into x
uff applies rule (transition rules)
uuffG(xd-x) u is control action x is
predicted world state xd is desired world
state uff is the feedforward control plan
31
Nodes are made from Modules
G is feedback
xd is from above
actions
sensors
transform into x
uff applies rule (transition rules)
if BALL, move toward centroid if NOT BALL, turn
clockwise (feedback determines how fast)
32
Nodes are made from Modules
G is feedback
xd is from above
actions
sensors
transform into x
uff applies rule (transition rules)
when to stop? how far is far enough? what about
noise/fuzzy ball? sensor noise, actuator
error, rigid models
if BALL, move toward centroid if NOT BALL, turn
clockwise (feedback determines how fast)
33
Advantages of Hierarchies
  • Albus and Mystel 01
  • Natural way to organize
  • Not intrinsically rigid
  • Not intrinsically inefficient
  • not the same as centralized planning
  • priorities and goals are clear, therefore
    efficient

34
Summary RCS
  • hierarchy with node structure at each level
  • have operator interface (in theory)
  • nodes consist of
  • Sensory Processing
  • World Model
  • Behavior Generation
  • Value Judgment
  • top-down, plan for a particular horizon
  • control theoretic

35
Evaluating the Two Architectures
  • support for modularity
  • decomposition by functionality
  • niche targetability
  • good, both have been used for apps like vehicle
    guidance, mining equipment
  • ease of portability to other domains
  • unclear, not sure if code could be reusedlots of
    rewriting on previous apps
  • robustness
  • RCA simulates plans in advance, but not sure what
    it would do with sensor or mechanical failures,
    etc.

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
36
Hierarchical Review
  • Describe the hierarchical paradigm in terms of
    the three robot primitives
  • Describe sensing in the hierarchical paradigm
  • What is STRIPS?
  • What is the closed world assumption?
  • What are preconditions?
  • What is the frame problem?
  • What are two representative architectures?
  • What is the NHC decomposition?

Organization -SPA -global Strips -Shakey Rep.
Arch. -evaluation -NHC -RCA Summary
Write a Comment
User Comments (0)
About PowerShow.com