Final Project Review Overview Sankt Augustin, February 15th, 2008 - - PowerPoint PPT Presentation

▶
final project review overview
SMART_READER_LITE
LIVE PREVIEW

Final Project Review Overview Sankt Augustin, February 15th, 2008 - - PowerPoint PPT Presentation

Aug 16, 2022 499 likes •1.08k views

Final Project Review Overview Sankt Augustin, February 15th, 2008 Erich Rome, Adaptive Reflective Teams Department ... and all others of the MACS consortium MACS Project Overview Reporting Period 3 1. MACS Approach MACS Approach 1.

slide-1
SLIDE 1

Final Project Review Overview

Erich Rome, Adaptive Reflective Teams Department

... and all others of the MACS consortium Sankt Augustin, February 15th, 2008

slide-2
SLIDE 2

MACS_Y3_Final_Review_Overview.ppt

2 FP6-004381-MACS

MACS Project Overview – Reporting Period 3

  • 1. Work Performed: WP overview, WP1&6 details
  • 1. MACS Approach

MACS Approach

  • 2. Work in Period 3

Work in Period 3

  • 3. Conclusion

Conclusion

  • 1. Introduction
  • 2. MACS Approach to Affordance-based Robot Control
slide-3
SLIDE 3

MACS_Y3_Final_Review_Overview.ppt

3 FP6-004381-MACS

MACS Project Overview – Reporting Period 3

  • 1. Work Performed:

Achievements, WP overview, WP1&6 details

  • 1. MACS Approach

MACS Approach

  • 2. Work in Period 3

Work in Period 3

  • 3. Conclusion

Conclusion

  • 1. Introduction
  • 2. MACS Approach to Affordance-based Robot Control
slide-4
SLIDE 4

MACS_Y3_Final_Review_Overview.ppt

4 FP6-004381-MACS

1.1 Introduction

Cognitive abilities:

 Using known (and unknown) objects in a variety of ways  Finding alternative solutions for a given task

➜ Ability to improvise; practical aspect of intelligence

 Exploiting specific interaction possibilities that the

environment offers humans

Source: www.flickr.com

J.J. Gibson: Affordances

slide-5
SLIDE 5

MACS_Y3_Final_Review_Overview.ppt

5 FP6-004381-MACS

1.1 Introduction

Research Questions in MACS:

 How can we design a “cognitive” mobile robot system

with manipulation capabilities that can, e.g.,

 find alternative solutions for a given task,  interact with known and unknown objects in a

meaningful and goal-directed way, and

 uses perception methods that are tailored for its tasks

and its action capabilities, i.e. that are grounded in its actions?

 Valid approach: Draw inspiration from Cognitive Science

➜ Affordances

slide-6
SLIDE 6

MACS_Y3_Final_Review_Overview.ppt

6 FP6-004381-MACS

1.1 Introduction – Gibson’s affordances

Characterizing affordances

 The “picked-up” visual information includes

visible functions (or utilities) of an object / thing / entity)

 These functions can be described using abstract

features (related to physical properties of the animal)

 The same object / thing / entity can offer different

functions for different animals

slide-7
SLIDE 7

MACS_Y3_Final_Review_Overview.ppt

7 FP6-004381-MACS

1.1 Introduction – Gibson’s affordances

Some conclusions:

 The characteristic feature set related to a function

works like a matched filter across various object categories.

 The ability to perceive affordances, i.e., to perceive

functions of entities in the world, enables more possibilities for action: An animal (or agent) could even guess what to do with entities that it has never before perceived.

sit! sit! sit! sit!

Affordance “sitable” offered by various entities in the environment

slide-8
SLIDE 8

MACS_Y3_Final_Review_Overview.ppt

8 FP6-004381-MACS

1.2 MACS Approach

Project goals:

 Explore how affordances can be used for robot control  Perception, learning, representation and goal-oriented

use of affordances

 Realisation of a complete affordance-inspired control

system

 Experimental evaluation  Proof of concept with a simulated and a real robot in a

demonstrator scenario

slide-9
SLIDE 9

MACS_Y3_Final_Review_Overview.ppt

9 FP6-004381-MACS

1.2 MACS Approach

MACS stance:

 Using affordances for deliberative control of a robotic

system requires a representation

 An explicit representation of affordances benefits from

a formalization of the affordance notion.

 Affordances are modelled as relations between the

abilities of an agent and features in the environment (based on proposals of Stoffregen, 2000ff, and Chemero, 2003ff)

slide-10
SLIDE 10

MACS_Y3_Final_Review_Overview.ppt

10 FP6-004381-MACS

1.2 MACS Approach – Formalisation

Entity: Perception aspect of the affordance. Relevant cues of the environment that provide support for the affordance. Behavior: Action aspect of the affordance in a robotic agent sense. Outcome: Outcome (or effect) of the agent’s acting upon an affordance (of applying the behavior)

Affordance Relation: (Entity, Behavior, Outcome)

Sahin et al. (2007): Adaptive Behavior 15(4).

Entity Outcome Behavior environment agent

slide-11
SLIDE 11

MACS_Y3_Final_Review_Overview.ppt

11 FP6-004381-MACS

1.2 MACS Approach – Formalisation

Agent affordance:

 An agent must be able to perceive an agent affordance.  “Entity” is usually not necessarily equal to “object”.

slide-12
SLIDE 12

MACS_Y3_Final_Review_Overview.ppt

12 FP6-004381-MACS

1.2 MACS Approach – Equivalences

Lift

Entity equivalence: A behavior applied on different entities produces the same outcome. {entity}: liftable entity characteristics, common, e.g., to blue and black can

Black can Entity Lifted Blue can

({entity}, behavior, outcome)

Sahin et al. (2007): Adaptive Behavior 15(4)

Liftability affordance: Lift Entity Lifted

slide-13
SLIDE 13

MACS_Y3_Final_Review_Overview.ppt

13 FP6-004381-MACS

1.2 MACS Approach – Affordance Representation

Definition:

 (Agent) Affordance Representation: data structure (cue descriptor, behavior descriptor, outcome descriptor), where

  • cue descriptor: entity representation, containing pairs of

attributes and associated value ranges,

  • behavior descriptor: reference to a robot behavior – reactive
  • r high-level –, plus an optional set of behavior parameters,
  • utcome descriptor: analogous to cue descriptor

Rome et al. (2006): D2.2.2

slide-14
SLIDE 14

MACS_Y3_Final_Review_Overview.ppt

14 FP6-004381-MACS

1.2 MACS Approach – Representation

Cue descriptor:

 The cue descriptor will contain acquired information

that is characteristic for the agent affordance at hand.

 Once learned, the cue descriptor can be used to perceive

an affordance. (➜ (matched) filter)

 The agent may refine the cue descriptor later when it

makes new experiences.

slide-15
SLIDE 15

MACS_Y3_Final_Review_Overview.ppt

15 FP6-004381-MACS

1.2 MACS Approach – Robot KURT3D

Robotic agent KURT3D:

Sensors:

  • 2 cameras
  • 3D Laser scanner
  • Distance transducers
  • weight sensor
  • more …

Actuators:

  • 6 wheels, 2 drive motors
  • 3-DOF crane with &

electromagnetic gripper

  • R. Breithaupt et al.

(FhG/AIS)

slide-16
SLIDE 16

MACS_Y3_Final_Review_Overview.ppt

16 FP6-004381-MACS

1.2 MACS Approach – Demonstrator Scenario

Demonstrator scenario:

 Two separated regions (“rooms”)  Sliding door, operated by switch  Switch triggerable by weight (adjustable)  Test objects:  Cylinders, spheres, boxes  Different tops, sizes, weights, color

combinations

Real MACSim

slide-17
SLIDE 17

MACS_Y3_Final_Review_Overview.ppt

17 FP6-004381-MACS

1.2 MACS Approach – Basic Skills

Basic perception and action capabilities (skills):

 Visual attention (bottom-up and top-down)  Feature detectors (single sensor, multi-sensor) as

Computational Units

 Pre-processing scanner data (free space, obstacles)  Roaming (uses Drive, Brake, Turn)  Approach_pose  Push, Lift, Drop, Carry, Stack, ...  Remote_control (for manual teaching) Behavior system

Basic skills

Perception module

Feature detectors

  • C. Lörken et al.

(FhG/IAIS – UOS) Paletta, Fritz, May, Ugur et al. (JR_DIB, FhG/IAIS, METU-KOVAN)

slide-18
SLIDE 18

MACS_Y3_Final_Review_Overview.ppt

18 FP6-004381-MACS

1.2 MACS Approach – Agent Affordances

Initial set of agent affordances:

 traversable

(offered by free space)

 pass(-through)-able

(region between things)

 push-able

(thing)

 lift-able

(thing)

 place-able

(offered by region) New and refined:

 switch-trigger-able

(lift-able + place-able)

 removable-from-switch

(lift-able + place-able)

 traversable

(free space + push-able)

slide-19
SLIDE 19

MACS_Y3_Final_Review_Overview.ppt

19 FP6-004381-MACS

1.2 MACS Approach – Characterizing Affordances

Characterizing some affordances:

 Lift-able characterization:  size limited by scanner resolution and FOV  magnetizable  limited weight (< 1kg)  flat top  color, shape may vary in limits of perceptual abilities

slide-20
SLIDE 20

MACS_Y3_Final_Review_Overview.ppt

20 FP6-004381-MACS

1.2 MACS Approach – Exploration and Application

Application Exploration Modes of operation:

 Simple pre-coded behaviors (basic skills)  Acquiring affordance knowledge (exploration)  autonomous or  controlled by human  Goal-directed (plan-based) affordance usage (application)

Simple pre- coded behaviors Interact with the environment Discover general relations Use relations in goal-directed behavior

slide-21
SLIDE 21

MACS_Y3_Final_Review_Overview.ppt

21 FP6-004381-MACS

1.2 MACS Approach – Bottom-up Perception

Bottom-up perception: Applying feature detectors

 Feature detectors (computational units):

 Saliencies in camera, range and remission images  Color blobs  SIFT categories  Blob size ratios  Behavior activation  Weight sensor  …

 Find characteristic subset

  • L. Paletta, G. Fritz

Joanneum Research Cue Cue SIFT regions Histogram categories

slide-22
SLIDE 22

MACS_Y3_Final_Review_Overview.ppt

22 FP6-004381-MACS

1.2 MACS Approach – Knowledge Acquisition

Acquiring knowledge on affordance relations

Performing (pre-coded) behavior (B)

Perceiving own actions (➜ learning)

Gaining knowledge about cues (C) (➜ learning)

Gaining knowledge about outcomes (O) (➜ learning)

Building C-B-O Affordance Representation Repository Behavior system Perception module

Feature detectors

Learning module Bottom-up acquisition

  • Aff. Repres. Repository

Dorffner, Kintzler, Irran OFAI

slide-23
SLIDE 23

MACS_Y3_Final_Review_Overview.ppt

23 FP6-004381-MACS

1.2 MACS Approach – Knowledge Acquisition

Training lifting:

  • Scan for cue
  • Steer crane arm
  • Lift
  • Scan for outcome / effect

Simulator:

  • MACSim (METU-KOVAN)
  • ODE based

Emre Ugur, METU-KOVAN

slide-24
SLIDE 24

MACS_Y3_Final_Review_Overview.ppt

24 FP6-004381-MACS

1.2 MACS Approach – Goal-oriented Application

Using knowledge (“top down”):

 “Mission” phase: Solve a given task

 Create a plan using affordance relations / representations:

A plan is a partially ordered set of operators, implemented by behaviors, with affordances in their preconditions

 Use cue information to purposively “look” for relevant cues

(➜ affordance cueing)

 Use equivalent descriptors if cue is not found (Execution control)

(➜ exploiting equivalences)

 Approach entity  Perform behavior  Check outcome

slide-25
SLIDE 25

MACS_Y3_Final_Review_Overview.ppt

25 FP6-004381-MACS

1.2 MACS Approach – Affordance Representation Repository

Using the Affordance Representation Repository

 Return (cues, behavior) tuples that lead to a given outcome

(➜ affordance equivalence)

 Return (behavior, outcome) tuples that are related to a given entity

(➜ entity equivalence)

 Return (cues, outcome) tuples that are related to a given action

(➜ behavior equivalence)

Tech report: Affordance Representation Tech report: Affordance Representation Repositors epositors (OFAI) OFAI)

Behavior system Perception module

Feature detectors

Learning module

  • Aff. Repres. Repository

Representation requests Execution module

Dorffner, Kintzler, Irran OFAI

slide-26
SLIDE 26

MACS_Y3_Final_Review_Overview.ppt

26 FP6-004381-MACS

1.2 MACS Approach – Top-down Perception

Top-down perception: Affordance cueing lift-able

  • L. Paletta, G. Fritz

Joanneum Research

Lift-able with flat top region Local feature extraction & classification Histogram: Peak for rectang. feat. Prediction using decision tree KLT based feature tracker for regions, verifying outcome

slide-27
SLIDE 27

MACS_Y3_Final_Review_Overview.ppt

27 FP6-004381-MACS

1.2 MACS Approach – Planning

Experimental implementation of the planner

 Domain: MACS scenario (simulated or physical)  “Affordance Map”: Room regions  Domain representation: PDDL  Planner: FF (Fast Forward)  Execution: Behaviors in MACS architecture

Affordances in scenario: passable, liftable, switch-triggerable, affords-removing-from-switch

Hertzberg, Lörken UOS

slide-28
SLIDE 28

MACS_Y3_Final_Review_Overview.ppt

28 FP6-004381-MACS

1.2 MACS Approach – Architecture v1

slide-29
SLIDE 29

MACS_Y3_Final_Review_Overview.ppt

29 FP6-004381-MACS

1.2 MACS Approach – Architecture v2

slide-30
SLIDE 30

MACS_Y3_Final_Review_Overview.ppt

30 FP6-004381-MACS

  • 2. MACS Project Overview – Reporting Period 3
  • 1. Work Performed:

Achievements, WP overview, WP1&6 details

  • 1. Introduction
  • 2. MACS Approach to Affordance-based Robot Control
  • 1. MACS Approach

MACS Approach

  • 2. Work in Period 3

Work in Period 3

  • 3. Conclusion

Conclusion

slide-31
SLIDE 31

MACS_Y3_Final_Review_Overview.ppt

31 FP6-004381-MACS

2.1 MACS – General Achievements

 Affordance-inspired Robot Control Architecture with

explicit affordance support in all parts

 Formalisations of agent affordances  Methods for perception and learning of affordances  Methods for planning and plan execution using

affordances

 Physics-based Simulator MACSim  Physical robot KURT3D  Dissemination

slide-32
SLIDE 32

MACS_Y3_Final_Review_Overview.ppt

32 FP6-004381-MACS

2.1 MACS – Work packages

 WP0 – Management  WP1 – Infrastructure  WP2 – Architecture  WP3 – Perception of Affordances  WP4 – Representation of Affordances / Deliberation  WP5 – Learning of Affordances  WP6 – Proof of Concept and Dissemination

slide-33
SLIDE 33

MACS_Y3_Final_Review_Overview.ppt

33 FP6-004381-MACS

2.1 MACS – WP 0 Management

Overview of Management achievements:

 Implementing reviewer recommendations (Periods 1+2)  Period 3: reporting and review preparations  Financial management, reporting, coordination  Submission of deliverables, project meetings  Work plan changes, 2nd EC contract amendment  Integrating the new partner UOS  Finalizing the project

slide-34
SLIDE 34

MACS_Y3_Final_Review_Overview.ppt

34 FP6-004381-MACS

2.1 MACS – WP 0 Management: Meetings

 3 plenary meetings

(Sankt Augustin, Altaussee, Osnabrück)

 3 managerial board meetings  12 technical meetings

(Sankt Augustin (8), Vienna, Ankara, Graz, Osnabrück)

slide-35
SLIDE 35

MACS_Y3_Final_Review_Overview.ppt

35 FP6-004381-MACS

2.1 MACS – WP 0 Management

Management deliverables:

 Deliverables in Period 3: Periodic Reports: D0.1.6 Interim Activity Report 3 D0.1.7a Periodic Activity Report 3 D0.2.3a Financial Information Period 3 D0.2.3c Periodic Management Report 3 Final Reports: D0.1.7b Final Activity Report D0.2.3d Final Management Report D0.2.3b Final report on the distribution of the Community‘s contribution

slide-36
SLIDE 36

MACS_Y3_Final_Review_Overview.ppt

36 FP6-004381-MACS

2.1 MACS – WP 1 Infrastructure

Behind the scenes:

What does it take to implement a project like MACS in terms of organisational and technical infrastructure?

Maintaining development server gibson

Keeping the project web site up-to-date

Keeping the document server DITO up-to-date

Maintaining the robot platforms and their components

Constructing and revising the robot’s crane arm

Constructing the demonstrator arena and test objects

Mechanical and electrical workshop, qualified staff

Providing a lab room for years

slide-37
SLIDE 37

MACS_Y3_Final_Review_Overview.ppt

37 FP6-004381-MACS

2.1 MACS – WP 1 Infrastructure

Overview of Infrastructure achievements:

 Computing infrastructure:

Maintaining many stationary and mobile computers

Setting up a distributed control system (RCS) for parallel execution of robot control routines on several computers, multiple processor kernels and GPUs

Performing stress tests, design software distribution

Find innovative solutions like GP-GPU implementations of resource hungry algorithms

Make the modules from different developers work in the integrated distributed system

GPU: ATI HD2400XT

slide-38
SLIDE 38

MACS_Y3_Final_Review_Overview.ppt

38 FP6-004381-MACS

2.1 MACS – WP 2 Affordance-based Architecture

Overview of Architecture achievements:

 Finalization of all modules, some simplified  Behavior system completed (UOS, FhG/AIS)  Execution Control completed (METU-KOVAN, UOS)  Integration of most of the modules  Architecture-related publications: 

Dagstuhl Springer LNAI 4760

Informatiktage 2007

ICRA 2008, EpiRob 2007

AAAI 2008

MACS control architecture diagram

slide-39
SLIDE 39

MACS_Y3_Final_Review_Overview.ppt

39 FP6-004381-MACS

2.1 MACS – WP 2 Affordance-based Architecture

Overview of Architecture achievements:

 Deliverables in Period 3: D2.3.1 Implementation of the affordance-based control architecture D2.3.2 A specification for a propositional planner and its interface to the MACS Execution Control Module

slide-40
SLIDE 40

MACS_Y3_Final_Review_Overview.ppt

40 FP6-004381-MACS

2.1 MACS – WP 3 Perception of Affordances

Overview of Perception achievements:

 Framework for the definition of affordance cues  Learning of affordance cues (~WP5)  2D and 3D affordance cueing  Real-time version of visual attention implemented on GPU

(➜ Demo)

 Improved stabilized triangulation method for exploration  Deliverables in Period 3:

D3.3.2 Sensorimotor decision making and affordance recognition V1+2

Closed-loop sensory-motor cueing Affordance cueing

slide-41
SLIDE 41

MACS_Y3_Final_Review_Overview.ppt

41 FP6-004381-MACS

2.1 MACS – WP 3 Perception of Affordances

Implementation:

 Visual attention implemented as a curiosity drive  IROS 2007  Visual recognition of affordance cues both in simulated

and real-world scenario

 ECVP 2006, SAB 2006, IROS 2006  Sensory-motor recognition of affordance cues  ICDL 2007, Dagstuhl Springer LNAI 4760  Exploitation of Laser-based depth imagery  Extension of MACS perception toolbox

slide-42
SLIDE 42

MACS_Y3_Final_Review_Overview.ppt

42 FP6-004381-MACS

2.1 MACS – WP 4 Representation for Deliberation

Overview of Representation achievements:

 WP lead taken over by new partner UOS  Specification of a planner for using affordance

representations

 Implementation of the planner for using affordance

representations

 Dagstuhl Springer LNAI 4760  Concept for grounding planning operators by affordances  CogSys 2008  Testing and using the planner both in simulation and on the

real robot

slide-43
SLIDE 43

MACS_Y3_Final_Review_Overview.ppt

43 FP6-004381-MACS

2.1 MACS – WP 4 Representation for Deliberation

Overview of Representation achievements:

 Second formalisation completed and published  Adaptive Behavior 15(4)  Deliverables in Period 3:

D4.3.3 A software prototype of the propositional MACS planning module D4.4.1 A software prototype for an affordance monitoring module with empirical testing using various MACS robotics platforms D4.4.3 An evaluation of the MACS planning module in the context of the MACS architecture D4.4.4 Submission of a conference or journal article describing the results of D4.3.3 and D4.4.3

Lift Liftability Black can Entity Lifted Blue can ({entity}, behavior, outcome)

slide-44
SLIDE 44

MACS_Y3_Final_Review_Overview.ppt

44 FP6-004381-MACS

2.1 MACS – WP 5 Learning of Affordances

Overview of Learning achievements:

 Reinforcement learning of affordance cues

 ECVP 2006, IROS 2006, AAPR 2007

 SVM-based supervised learning

 ICDL 2007b

 Three-stage developmental learning method for

acquiring knowledge on affordances

 ICDL 2007a, KI 2007, Dagstuhl Springer LNAI 4760

 Concept paper: Outlook towards affordance usage

  • bservation and imitation (D5.4.5)
slide-45
SLIDE 45

MACS_Y3_Final_Review_Overview.ppt

45 FP6-004381-MACS

2.1 MACS – WP 5 Learning of Affordances

Overview of Learning achievements:

 Deliverables in Period 3:

D5.3.3 Robot protoype learning affordances through self-experience V1+2 D5.4.5 Outlook towards affordance usage observation and imitation

slide-46
SLIDE 46

MACS_Y3_Final_Review_Overview.ppt

46 FP6-004381-MACS

2.1 MACS – WP 6 Dissemination

Industry Day, Jan 29, 2008:

 Cognitive Systems Industry Day, Schloss Birlinghoven,

Sankt Augustin, Germany

 Implemented as a Network Action from euCognition  Organized by FhG/AIS  47 attendees, 23 from industry  Projects: JAST, CoSy, RobotCub, MACS, URUS, SPARK  Welcome address (Prof. Th. Christaller), EU keynote (Cécile

Huet), poster session, MACS presentation (E. Rome)

 Presentations available at ftp server  Web site: http://www.macs-eu.org/csid

slide-47
SLIDE 47

MACS_Y3_Final_Review_Overview.ppt

47 FP6-004381-MACS

2.1 MACS – WP 6 Dissemination

Attention in Cognitive Systems Jan 8, 2007:

 International Workshop, co-located with AAAI 07 in

Hyderabad, India

 Organized by L. Paletta (JR_DIB) and E. Rome (FhG/AIS)  Extended post-proceedings appeared as Springer LNAI 4840

Paletta, Rome (eds.): Attention in Cognitive Systems

 Web site: http://dib.joanneum.at/wapcv2007/  Paper: Paletta & Fritz (JR_DIB)

slide-48
SLIDE 48

MACS_Y3_Final_Review_Overview.ppt

48 FP6-004381-MACS

2.1 MACS – WP 6 Dissemination

Künstliche Intelligenz 2007, Sep 10–13, 2007:

 German Conference on Artificial Intelligence, KI 2007,

University of Osnabrück, Germany

 Organized by J. Hertzberg (UOS), M. Beetz (TUM), R. Englert

(Telekom Laboratories)

 Proceedings appeared as Advances in Artificial Intelligence,

Springer LNAI 4667, Hertzberg, Beetz, Englert (eds.)

 Web site: http://www.ki2007.uos.de/  Papers from UOS and JR_DIB

slide-49
SLIDE 49

MACS_Y3_Final_Review_Overview.ppt

49 FP6-004381-MACS

2.1 MACS – WP 6 – Publications

Publications by Reporting Period

1 1 3 12 15 2 6 8 2 2 4 2 2 5 10 15 20 25 30 35 P1 P2 P3 Total Poster Theses Workshop Proceedings Peer-reviewed Conferences and Workshops Peer-reviewed Journals

slide-50
SLIDE 50

MACS_Y3_Final_Review_Overview.ppt

50 FP6-004381-MACS

2.1 MACS – WP 6 – Publications

Journals (peer-reviewed):

 Adaptive Behavior

Conferences (peer-reviewed):

 IROS 2006, IEEE/RSJ International Conference on Intelligent Robots and Systems  ICRA 2007, IEEE International Conference on Robotics and Automation  IASTED International Conference on Robotics and Applications 2007  2 x IROS 2007, IEEE/RSJ International Conference on Intelligent Robots and Systems  ICRA 2008, IEEE International Conference on Robotics and Automation  SAB 2006, Simulation of Adaptive Behavior (From Animals to Animats)  2x ICDL 2007, International Conference on Developmental Learning  ECVP06 29th European Conference on Visual Perception  AAPR / ÖAGM, 31st Workshop of the Austrian Association for Pattern Recognition  EpiRob 07, International Conference on Epigenetic Robotics  EMCSR 06, 18th European Meetings on Cybernetics and Systems Research  Informatiktage 2007, German Informatics Congress  KI 2007, German Conference on Artificial Intelligence

Legend:

  • Robotics
  • Learning
  • Perception
  • Multidisciplinary / CogSci
  • AI / Informatics
slide-51
SLIDE 51

MACS_Y3_Final_Review_Overview.ppt

51 FP6-004381-MACS

2.1 MACS – WP 6 – Publications

Peer-reviewed, accepted:

 Dogar, M.R.; Sahin, E.; Ugur, E. and Cakmak, M.: “Using Learned Affordances for Robotic Behavior Development.”

Accepted for the IEEE International Conference on Robotics and Automation, ICRA 2008.

Peer-reviewed, submitted:

 E. Rome; Paletta, L.; Doherty, P.; Sahin, E. and Dorffner, G.: “Affordance-related Robotics Research – A Survey.”

submitted to Adaptive Behavior.

 Lörken, C. and Hertzberg, J.: “Grounding planning operators by affordances.” submitted to CogSys 2008.  Ugur, E.; Sahin, E. and Dogar, M.R.: “Planning with Learned Object Affordances.“ submitted to AAAI 2008.

Edited books:

 Paletta, L. and Rome, E. (eds): Attention in Cognitive Systems. Springer-Verlag, Berlin, LNAI 4840, January 2008.  Rome, E.; Hertzberg, J. and Dorffner, G. (eds): “Towards Affordance-based Robot Control.” Proceedings of Dagstuhl

Seminar 06231. LNAI 4760, Springer-Verlag, Berlin, February 2008.

slide-52
SLIDE 52

MACS_Y3_Final_Review_Overview.ppt

52 FP6-004381-MACS

2.1 MACS – WP 6 – Publications

Book contributions:

 Paletta, L. and Fritz, G. “Reinforcement Learning for Decision Making in Sequential Attention.” In: Attention in

Cognitive Systems, LNAI 4840, pp. 293-306, Springer-Verlag, Berlin, Germany, Paletta, L., and Rome, E. (eds.), 2008.

 Paletta, L. and Fritz, G.: “Reinforcement Learning of Predictive Features in Affordance Perception.” In: Towards

Affordance-based Robot Control, Proceedings of Dagstuhl Seminar 06231, Springer-Verlag, Berlin, LNAI 4760, Rome, E., Hertzberg, J. and Dorffner, G. (eds.), February 2008.

 Hertzberg, J.; Lingemann, K.; Lörken, C.; Nüchter, A. and Stiene, S. “Does it Help a Robot Navigate to Call

Navigability an Affordance?” In: Towards Affordance-based Robot Control, Springer-Verlag, Berlin, LNAI 4760, Rome, E., Hertzberg, J. and Dorffner, G. (eds.), February 2008.

 Irran, J.; Kintzler, F. and Dorffner, G.: “Learning of Interaction Possibilities. ”In: Towards Affordance-based Robot

Control, Springer-Verlag, Berlin, LNAI 4760, Rome, E., Hertzberg, J. and Dorffner, G. (eds.), February 2008.

 Rome, E.; Paletta, L.; Sahin, E.; Dorffner, G.; Hertzberg, J.; Fritz, G.; Irran, J.; Kintzler, F.; Lörken, C.; May, S.; Ugur, E.;

Breithaupt, R.: “The MACS project: An approach to affordance-based robot control.” In: Towards Affordance-based Robot Control, Springer-Verlag, Berlin, LNAI 4760, pp. 173-210, Rome, E., Hertzberg, J. and Dorffner, G. (eds.), February 2008.

Online articles:

 E. Rome, J. Hertzberg, G. Dorffner and P. Doherty (2006): Towards Affordance-based Robot Control – Executive

Summary of the Dagstuhl Seminar 06231. In: Online proceedings of Dagstuhl Seminar 06231.

 E. Rome, J. Hertzberg, G. Dorffner, P. Doherty et al. (2006): Towards Affordance-based Robot Control – Abstracts

Collection of the Dagstuhl Seminar 06231. In: Online proceedings of Dagstuhl Seminar 06231.

slide-53
SLIDE 53

MACS_Y3_Final_Review_Overview.ppt

53 FP6-004381-MACS

2.1 MACS – WP 6 – Publications

Theses:

 Lörken, C.: Introducing Affordances into Robot Task Execution. Masters Thesis, University of Osnabrück, Institute for

Cognitive Science, 2006. Available online at PICS, Volume 2-2007.

 Ugur, E.: “Direct Perception of Traversability Affordance on Range Images Through Learning on a Mobile Robot.”

M.Sc. Thesis, Middle East Technical University, Kovan Laboratory, 2006.

 Cakmak, M.: “Robot Planning based on Learned Affordances.” M.Sc. Thesis, Middle East Technical University, Kovan

Laboratory, July 2007.

 Dogar, M.R.: “Using Learned Learned Affordances for Robot Behavior Development.” M.Sc. Thesis, Middle East

Technical University, Kovan Laboratory, September 2007.

Affordance research bibliography:

Collected by all partners, online provided via METU-KOVAN’s WebBibTeX: http://www.kovan.ceng.metu.edu.tr/wbt/

Web Site:

Up-to-date web site with complete coverage of project activities: http://www.macs-eu.org/

Deliverables:

D6.5.2 Industry Day D4.4.4 Submission of a conference or journal article describing the results of D4.3.3 and D4.4.3 (CogSys 08 submission)

slide-54
SLIDE 54

MACS_Y3_Final_Review_Overview.ppt

54 FP6-004381-MACS

2.1 MACS – WP 6 – Dissemination

More dissemination achievements:

 Jan 24, 2007: Karlsruhe, Germany

VDI-GMA Fachausschusssitzung 4.13 Robotersysteme Erich Rome: Invited MACS Approach Talk

 July 16, 2007: Edinburgh, UK

Seminar at University of Edinburgh, Institute for Perception, Action and Behavior Emre Ugur (METU-KOVAN): The learning and use of traversability affordance on a mobile robot

 12 Posters at various events

slide-55
SLIDE 55

MACS_Y3_Final_Review_Overview.ppt

55 FP6-004381-MACS

2.1 MACS – WP 6 – Proof of Concept

Overview of Proof of Concept achievements:

 Traversability experiments:

Traversability Maps (LiU-IDA, D4.3.1)

 D4.2.1+4.3.1

Extended traversability affordance (METU-KOVAN)

 ICDL 2007, ICRA 2007

3D range image based traversability (UOS), learning relevant features for perceptual economy

 IASTED RA 2007

slide-56
SLIDE 56

MACS_Y3_Final_Review_Overview.ppt

56 FP6-004381-MACS

2.1 MACS – WP 6 – Proof of Concept

Overview of Proof of Concept achievements:

 Liftability learning experiments in MACSim (METU-KOVAN)

 AAAI 2008

 Goal-directed usage of liftability affordance (UOS)

 CogSys 2008

 Affordance cueing (JR_DIB, FhG/AIS)

 IROS 2006, AAPR 2006, ECVP 2006, IROS 2007

 “Curiosity Drive” (FhG/AIS, METU-KOVAN)

 IROS 2007, ICDL 2007

 Deliverables in Period 3: D6.4.2 Report on experimental results in simulator D6.4.3 Report on experimental results in demonstrator

slide-57
SLIDE 57

MACS_Y3_Final_Review_Overview.ppt

57 FP6-004381-MACS

MACS Project Overview

Thank you for your attention!

Acknowledgments: FhG/AIS: R. Breithaupt, S. Frintrop , C. Hoffmann , M. Hülse, M. Klodt,

  • M. Magnusson, S. May, B.S. Müller, P. Müller, M. Pintea, D. Qiu, F.

Schönherr, I. Stratmann, H. Surmann, R. Worst Joanneum Research: G. Fritz, M. Kumar, L. Paletta, C. Seifert Linköpings Universitet: P. Doherty, F. Heintz, T. Merz, T. Persson, P. Rudol , B. Wingman, M. Wzorek METU-KOVAN: E. Sahin, M. Cakmak, M. Dogar, F. Nar, S. Eren, G. Ücoluk, E. Ugur, M. Yavas OFAI: G. Dorffner, J. Irran, F. Kintzler, A. Lewandowski, P. Pölz University of Osnabrück: A. Bartel, J. Hertzberg, C. Lörken, F. Meyer