1 / 55

Steps Toward Developing an Intelligent Robotics Course Kutztown University

Steps Toward Developing an Intelligent Robotics Course Kutztown University. PACISE 2011 April 9, 2011 Oskars J. Rieksts Jeffrey W. Minton. Acknowledgments – Funding. CS Department & LAS College TYCO Electronics Foundation Kutztown University Foundation

Download Presentation

Steps Toward Developing an Intelligent Robotics Course Kutztown University

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Steps Toward Developing an Intelligent Robotics CourseKutztown University PACISE 2011 April 9, 2011 Oskars J. Rieksts Jeffrey W. Minton

  2. Acknowledgments – Funding • CS Department & LAS College • TYCO Electronics Foundation • Kutztown University Foundation • KU Faculty Professional Development Committee • KU Undergraduate Research Committee • PASSHE Faculty Professional Development Council 2011 Kutztown University 2

  3. Motivations • Rodney Brooks • DARPA Grand Challenge • REU @ Auburn • Hardware-software synthesis • “Hardened” hardware • Central role of software 2011 Kutztown University 3

  4. Motivations • Applied A.I. • Applied cognition • Mobility • Assistive robotics • Robotic wheelchairs 2011 Kutztown University 4

  5. Outline • Early efforts • Hacking Roomba & other ventures • Research Experience for Undergraduates • Create and Mindstorms • Captain KURK • Trinity firefighting contest • Myro and Python • Reading-Berks Science Fair 2011 Kutztown University 5

  6. Coming into Focus • Emerging goals • Intelligence • Vision • Communcation • Philosophical issues 2011 Kutztown University 6

  7. Point of convergence • Course in intelligent robotics • Control • Vision • Communcation • Investigate cognitive issues 2011 Kutztown University 7

  8. Course design objectives • Robot is • Situated • Embodied • Shared environment • Human • Machine • Shared communication • Framework for investigation 2011 Kutztown University 8

  9. Robot is situated • Operates within an environment • Embedded in the world • Chief knowledge source: • Data stream drawn from environment • “The world is its own best representation” – Brooks 2011 Kutztown University 9

  10. Robot is embodied • Entity 1st, agent 2nd • Self reliant • Secretary of State model • Self-seated conceptual framework • World concepts grounded in sensor suite • Behavior set emanates from actuator suite • Decision apparatus grounded in sensor/actuator suites 2011 Kutztown University 10

  11. Robot is embodied • Eames: Design is a plan for arranging elements to accomplish a particular purpose • Elephants don’t play chess (Brooks) • “Mind” is to fit the body • Design on need to basis • Know • Think • Do 2011 Kutztown University 11

  12. Shared environment • Shared conceptual framework • Robot’s • Derivative of human’s • Simplified in structure • Shared factbase • Robot’s • Subset of human’s • Simplified ontology • Vision based 2011 Kutztown University 12

  13. Common communication framework • Grounded in • Shared • Environment • Conceptual framework • Ontology • Vision as main sensor • Intersecting language constructs 2011 Kutztown University 13

  14. Projected course topics/activities • Specialized robot control software • Robot control architecture • Basics of image processing • Specialized image processing software • Basics of communication theory • Key issues of cognitive robotics 2011 Kutztown University 14

  15. Slow, steady progress • Browning: • Reach should exceed grasp • Rapid prototyping • Iterative development 2011 Kutztown University 15

  16. Structure of the class • Four teams • Roles • Team leader • Designer • Coder • Document guru • Historian • Test designer • Test administrator • Hardware specialist 2011 Kutztown University 16

  17. PACT demonstration • Pennsylvania Association of Council of Trustees • Very early in learning curve • Navigation within a “corral” • Well received • Gateway navigation • Bull fighting robot • PR for robotics @ PASSHE 2011 Kutztown University 17

  18. Jeff – gateway navigation 2011 Kutztown University 18

  19. Investigate cognitive issues • Sight and touch • George Stratton • Spatial harmony of sight & touch • Space • Benjamin Kuipers • Semantic spatial hierarchy • Metrical • Topological • Hybrid 2011 Kutztown University 19

  20. Investigate cognitive issues • Language and meaning • Stevan Harnad • Symbol grounding problem • Language and shared space • Amichai Kronfeld • Shared referent problem 2011 Kutztown University 20

  21. Investigate cognitive issues • Embodied cognition • Randall Beer • Situated, embodied, dynamical • Principle of interactivism • Mark Bickhard • Emergence of representational content 2011 21 Kutztown University

  22. Illustration – Shared Referent Problem • From Kronfeld • Steve & wife Bev are at a party • Steve: Bob & his wife are certainly enjoying the party • Bev: Sally is not his wife! • Note • Steve & Bev share a referent, Sally • Despite misidentification no communication problems arise! Kutztown University 2011 22

  23. The KLO Triad • For communication involving referents • For each communicant – this triad • Object (the referent) • KB – representation of object • Language – reference to object • For successful communication • Referents must match • For HMC (Human-machine communication) • Human & machine KBs do not match 2011 Kutztown University 23

  24. Establishing co-refrence • Per Frege’s distinction • Extensional approach • Pointing or equivalent • Intensional approach • Language alone • HMC goal • Minimize extensional • Maximize intensional 2011 Kutztown University 24

  25. The Trinity Robot 2011 Kutztown University 25

  26. The Trinity Robot • VEX robotics kit for chasis • VEX included motors for locomotion • Sonar rangefinders • Web-cam • Arduino to control motors and sensors • Netbook to control Arduino and process images 2011 Kutztown University 26

  27. The Trinity Robot • Inaccurate motor control • Sonar signals bounce inside corners • Provide inaccurate measurements • Viewing angle of web-cam too small 2011 Kutztown University 27

  28. Brobi 2011 Kutztown University 28

  29. Brobi – hardware • Create by iRobot • Platform built onto Create cargo bay to accommodate equipment • Web-cam with increased viewing angle • IR rangefinder • IR light does not bounce like sound • Arduino 2011 Kutztown University 29

  30. Brobi – software • OpenCV and Python • Consultation – John Spletzer • Lehigh • Little Ben in Urban Challenge • MATLAB • Image processing • Create API • A programming language 2011 Kutztown University 30

  31. Image processing • “A picture is worth a thousand words” • need to extract discrete objects from images to identify them • K-Means clustering 2011 Kutztown University 31

  32. Goal: identify green ball 2011 Kutztown University 32

  33. K-Means Clustering • Cluster sets of data • Into user-defined number of segments • Number of segments referred to as K • Clusters defined by MEAN of all values in cluster 2011 Kutztown University 33

  34. K-Means Algorithm

  35. K-Means Example K-Means using k = 5 K-Means using k = 6

  36. Natural Language Processing • “Go to the green ball” • The meanings behind words must be inferred • “Go,” conceptually can represent many things • Take a turn in a game • The Chinese strategy game • Travel to a location • Determine the concept being referred to • Conceptual parsing 2011 Kutztown University 36

  37. Conceptual Parsing • Words are mapped to concepts • Concepts • Rules define set of related concepts • One concept may have many separate rule sets 2011 Kutztown University 37

  38. Concept Tree

  39. Assessment – Platform • Best hardware platform to date • Opens up many avenues of course development • Software for robot control systems • Robot control architectures • Image processing • Communication • Cognitive robotics issues 2011 Kutztown University 39

  40. Assessment – MATLAB • Can do all 3 things • Sensor/actuator interface with Lehigh API • Image acquisition & processing • Robot control system programming • Good IDE • Good documentation • 17 pdf files; 64 mB • Online documentation 2011 Kutztown University 40

  41. Assessment – MATLAB • Good IDE • Strong user community • Blogs • Discussion boards • Good code segments • Good programming language • Matrix optimized • Advance features, e.g., lambda, apply 2011 Kutztown University 41

  42. Assessment – Create • Useful • Stable • Ubiquitous • Negatives • Battery short lasting • Does not travel in straight line 2011 Kutztown University 42

  43. Assessment – Accessibility • Cost not prohibitive • Create – $130 • $220 with battery and charger • Netbook ~ $330 • Arduino ~ $30 • MATLAB • $900 for 10 seat license • $130 for student version with image acquisition 2011 Kutztown University 43

  44. Future directions • Software repository • Learning • Bayesian • Genetic algorithms • Neural networks, etc. • Control architecures • Behavior-based • Hybrid • Blackboard? 2011 Kutztown University 44

  45. Future directions • Experiment with other sensors • Touch {whiskers} • Odor • Sound {whistle, a la Sound of Music} • Heat • Odor • Light/brightness • Wireless interface • “Watson, come here. I need you” 2011 Kutztown University 45

  46. Future of robotics at KU • Strong student interest • Rejuvenated course • New syllabus • Approved • 400 level – a mixed blessing • Not yet scheduled • In rotation? • Politics 2011 Kutztown University 46

  47. Further Information • Jeff’s graduate thesis • Do You See What I'm Saying: Relating Language and Vision to Create Interaction Between Humans and Robots • Delves further into concepts discussed here • https://github.com/jeffminton/Thesis 2011 Kutztown University 47

  48. Questions? 2011 Kutztown University 48

  49. The End 2011 Kutztown University 49

  50. Extra slides 2011 Kutztown University 50

More Related