Plan-based Control in an Plan-based Control in an Affordance-based - - PowerPoint PPT Presentation

▶
plan based control in an plan based control in an
SMART_READER_LITE
LIVE PREVIEW

Plan-based Control in an Plan-based Control in an Affordance-based - - PowerPoint PPT Presentation

Feb 23, 2023 116 likes •348 views

Plan-based Control in an Plan-based Control in an Affordance-based Robot Control Affordance-based Robot Control Architecture Architecture Joachim Hertzberg, Christopher Lrken Institute for Informatics www.inf inf.uos. .uos.de/kbs/

slide-1
SLIDE 1

1

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Plan-based Control in an Plan-based Control in an Affordance-based Robot Control Affordance-based Robot Control Architecture Architecture

Joachim Hertzberg, Christopher Lörken Institute for Informatics www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Thanks to

Andreas Bartel, Kai Lingemann, Frank Meyer, Andreas Nüchter, Stefan Stiene

slide-2
SLIDE 2

2

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Planning & Execution in MACS Planning & Execution in MACS

Retry

slide-3
SLIDE 3

3

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

From the WP4 Objectives From the WP4 Objectives

  • Understand interplay sensor data – action usage – affordances

in robotics

  • Understand inter-relation affordances – symbol grounding in robotics
  • Specify generic representations & data structures supporting affordance

usage

  • Understand how affordances reified as data structures interact with signal-

to-symbol conversions, action structures, and qualitative representations of

  • bjects
  • Propose software module supporting a sensor data – dynamic object –

action execution – sensor usage loop on a robot

  • Study the relation affordances – planning for sensor use,

where affordances may be viewed as constraints on sensor use

  • Propose sensor planner operating dependently with the above
slide-4
SLIDE 4

4

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

  • Tailor (some)
  • perators after

afforded actions

  • Use affordances

as preconditions for “opportunistic” execution

  • Map perceived

affordances

  • Don’t distinguish

among function- ally equal objects

  • Reduce search

Ground (some) domain concepts in 〈c,b,o〉s

  • Focus attention
  • n affordances

serving current

  • perator;

disregard others

  • Search actively

for cues signaling focused affordances

Top-down influence on affordance usage

UOS UOS’ ’s s Contribution to WP4 Objectives Contribution to WP4 Objectives

Affordance-Based Robot Control Plan-Based Robot Control Use s-o-a AI planning technology

(PDDL language, FF planner)

Use the MACS modules

  • exec. control, behaviors,

affordance repository

slide-5
SLIDE 5

5

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Background: (Propositional) AI Planning Background: (Propositional) AI Planning

  • Dates back to STRIPS/SHAKEY tradition in AI
  • Modern algorithms (“neo-classical planning”) by orders of

magnitude faster, termination guaranteed

  • We re-used FF (Hoffmann/Nebel, 2000’s)
  • Well-understood on formal grounds
  • Situation is a set of ground facts
  • Operator pre/postconditions sets of ground literals
  • Plan is a partially ordered operator set
  • Decidable, NP-hard wrt. domain size
  • De-facto standard domain descr. language PDDL (+ variants)

…

slide-6
SLIDE 6

6

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

PDDL: PDDL: s-o-a s-o-a Domain Descriptions Domain Descriptions

  • Originally developed to facilitate the International Planning

Competition (IPC)

  • Codifies input syntax for specifying planning domains &

problems in terms of

  • predicates (proposition schemata)
  • actions
  • objects
  • start situation, goal propositions
  • optional requirements (typing, equality handling, …)
  • Various upgrades exist for enhancing expressivity (time, …)
slide-7
SLIDE 7

7

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Background: Robot Planning Background: Robot Planning

The plan is that part of the robot’s program, whose future execution the robot reasons about explicitly. [D. McDermott, 1992]

  • Dates back to STRIPS/SHAKEY tradition in AI
  • Various benefits for robot ctrl: Performance optimization (time,

robustness), communication, software engineering

  • Plan is just one source of information for robot control

(hybrid control architectures)

  • “Sense-Model-Plan-Act” (SMPA) loop is a straw man!
  • Plan format may vary; notion of planning may differ from

classical view (“adapting library plan stubs”)

☛ Autonomous execution matters ☛ Needs symbol grounding/object anchoring & action grounding

slide-8
SLIDE 8

8

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Example: MACS Arena Affordance Map Example: MACS Arena Affordance Map

Topological Region

(fuzzy boundaries)

Perceived Affordance

(arbitrarily often; possibly different cues) While being in some environment, log perceived affordance types per region! (1 entry / type / region)

slide-9
SLIDE 9

9

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

MACS DD Example: Predicates MACS DD Example: Predicates

(:types region switchRegion doorRegion room) (:predicates (robotAt ?region - region) (inRoom ?region - region ?room - room)) (hasLiftedSomething) (liftable ?region - region ) (switch-triggerable ?region - switchRegion) (passable ?startRegion ?targetRegion - region)) Model affordances by properties of the regions where they have been perceived

(no objects sneaking in through the backdoor)

slide-10
SLIDE 10

10

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

MACS DD Example: An Operator MACS DD Example: An Operator

(:action action lift :parameters parameters (?region - region) :precondition precondition (and (robotAt ?region) (liftable ?region) (not (hasLiftedSomething))) :effect effect (and (hasLiftedSomething) (not (liftable ?region)))) Grounded by localization Grounded by affordance (map) Grounded by “introspection” Side effect: delete liftability tag for

?region in aff. map

slide-11
SLIDE 11

11

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Execution: Grounding Operators Execution: Grounding Operators

  • Ground operators in behaviors ( hybrid architecture)
  • e.g., lift operator is implemented using:

DirectGoToPoseBehavior, 3DScanBehavior, ReachBehavior, PullBehavior, RaiseBehavior

  • Specialty induced by affordances:

If an operator corresponds to an afforded action, then grounding is provided by the b,o in 〈c,b,o〉!

  • Execution monitoring of afforded action means to

“go with the flow of affordance” (But only the selected one!)

slide-12
SLIDE 12

12

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Example: Grounding the Lift Operator Example: Grounding the Lift Operator

slide-13
SLIDE 13

13

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Opportunistic Execution Opportunistic Execution

[planning/execution] in tasks fraught with complexity and uncertainty might benefit from less of the discipline imposed by a top-down process. [B. & F. Hayes-Roth, 1979]

  • At operator execution, perception is primed to attend to cues

relevant for current operator execution

  • Any(!) entity affording what is needed may be used

(“lift something” vs “lift object O_17”)

  • Purely object-based representations handle poppycock

(“get me Glass_42 with water” instead of “get me a glass of water”)

☛ Using entities&affordances in addition

to(!) objects&properties appears to make a lot of sense!

slide-14
SLIDE 14

14

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

MACS Problem Description & Objects MACS Problem Description & Objects

(define (problem macs-prob) (:domain macs-example) (:objects rightRoom - room leftRoom - switchRoom region1_left region2_left region1_right - region switchRegion - switchRegion doorRegionLeft doorRegionRight - doorRegion ) (:init (inRoom region1_left leftRoom) (inRoom region2_left leftRoom) (inRoom switchRegion leftRoom) (inRoom doorRegionLeft leftRoom) (inRoom region1_right rightRoom) (inRoom doorRegionRight rightRoom) (robotAt region1_left) (liftable region1_left) (switch-triggerable switchRegion)) (:goal (robotAt doorregionright))) Types, predicates, actions Static Domain Features Dynamic “Fluents”

slide-15
SLIDE 15

15

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Complete Example (Simulator) Complete Example (Simulator)

The Plan (FF generated)

0: LIFT region1_left 1: CARRY region1_left switchregion 2: TRIGGER-SWITCH switchregion 3: APPROACH-REGION switchregion doorregionleft 4: CHANGE-ROOM doorregionleft doorregionright

The Goal

(:goal (robotAt doorregionright)))

slide-16
SLIDE 16

16

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Complete Example (Simulator) Complete Example (Simulator)

slide-17
SLIDE 17

17

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Execution Failure Execution Failure

  • Planned operator execution by afforded actions may

fail due to

  • Model error (affordance not present where in map)
  • Perception error (cue is there but gets overlooked;

affordance is perceived false-positively)

  • Handling error (afforded behavior fails)
  • Reaction inventory in plan-based control:

retry, replan, give up

  • Afforded actions may be retried by using different

affordance instance of the same type (perceived or looked up in map)

slide-18
SLIDE 18

18

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Execution Failure etc., Simulator Execution Failure etc., Simulator Expl Expl. .

slide-19
SLIDE 19

19

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Future Work Future Work

Examine interplay plan-based & affordance-based control

  • Continue/extend experiments

(real robot, opportunism, execution failure)

  • Examine more expressive plan language (time)
  • Interface with learning
  • Integrate individual objects
slide-20
SLIDE 20

20

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Summary of Contributions Summary of Contributions

 p-b ctrl helps a-b ctrl focus top-down on relevant affordances  a-b ctrl helps p-b ctrl ground actions and predicates  employed s-o-a propositional planner in robot ctrl  found operational model for opportunistic plan execution  found simple mechanism for handling anonymous entities/objects in propositional robot planning … and integrated it all in the MACS architecture The interplay between affordance-based and plan-based robot control has been explored for the first time

slide-21
SLIDE 21

21

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/

Publications Publications

  • 1. A. Bartel, F. Meyer, C. Sinke, T. Wiemann, A. Nüchter, K. Lingemann, and J. Hertzberg:

Real-Time Outdoor Trail Detection on a Mobile Robot. In: Proc. 13th IASTED Intl. Conf. Robotics and Applications, pp. 477–482, Würzburg, Germany, August 2007

  • 2. J. Hertzberg, K. Lingemann, C. Lörken, A. Nüchter, and S. Stiene (2007): Does it Help a

Robot Navigate to Call Navigability an Affordance? In: [5] pp.16–26, February 2008

  • 3. C. Lörken. Introducing affordances into robot task execution. In K.-U. Kühnberger, P.

König, and P. Ludewig (eds:), Publications of the Institute of Cognitive Science (PICS), vol. 2, 2007. University of Osnabrück, May 2007

  • 4. C. Lörken and J. Hertzberg: Grounding Planning Operators by Affordances. submitted for

publication, Nov. 2007

  • 5. E. Rome, J. Hertzberg, and G. Dorffner (eds): Towards Affordance-based Robot
  • Control. Proceedings of Dagstuhl Seminar 06231. Springer LNAI 4760, Berlin, February

2008

… during MACS funding 2007 and after; beyond deliverables

slide-22
SLIDE 22

22

www. www.inf inf.uos. .uos.de/kbs/ de/kbs/