b65ac089428cc323983bea5bc8503e46.ppt
- Количество слайдов: 42
REALITY SHOW OR TRUSTABLE IMPOSTORS ? Alex Garcia-Alonso University of the Basque Country /18
Outline • • Research environment Reality show or trustable impostor Recent research activities Your input & discussion Chapel Hill, Oct. 23 rd
Research environment • Basque Country – 4 Universities (90, 000 students) – 12 Tech research centers (2, 200 labor force) Chapel Hill, Oct. 23 rd
@ CEIT (1982 -1995) • Technological Research Center • App. Mechanics Dpt. : comp. graphics – FEM : pre- post-processing – Interactive 3 D visualization • Mechanisms analysis • Dynamics … – Collision analysis (RT mechanisms) … • Working for different companies – European Space Agency – MDI (Mechanical Dynamics Inc, Ann Arbor MI) – EADS … Chapel Hill, Oct. 23 rd
@ Univ. of the Basque Country (1996 - ) • Teaching !!! • Technological Research centers – CEIT (258 Labor staff) • University of Navarre – Ikerlan (230 Labor staff) • MCC Cooperative group 250 companies -- 83, 600 labor force • 44% Basque Country 37% Spain 19% international – Labein (356 Labor staff) – Vicomtech – EUVE … www. landersimulation. com • Spin-offs +VR www. simtechniques. com Chapel Hill, Oct. 23 rd
Research environment U. of the Basque Country U. of Navarre CEIT Vicomtech U. of Deusto Labein 100 miles diameter MCC – Ikerlan U. of Mondragon Bilbao San Sebastian Mondragon Vitoria EUVE Chapel Hill, Oct. 23 rd
SKILLS research network • • • • Multimodal Interfaces for CEIT capturing and Transfer of Skill Labein Aalborg University Commissariat à l'Energie Atomique - Laboratoire d'Intégration des Systèmes et des Technologies (CEA LIST) DLR - Institute of Robotics and Mechatronics Fraunhofer Institut für Graphische Datenverarbeitung HAPTION KUKA Roboter Gmb. H OMG plc PERCRO, - Italy [coordinator] Queen's University of Belfast Simonazzi S. p. A. Technion - Israel Institute of Technology UM 1 - University Montpellier-1 University of Tampere Chapel Hill, Oct. 23 rd
ENACTIVE research network • • • Labein Learn by doing CEIT Scuola Superiore Sant'anna, PERCRO Queen's University of Belfast Institut National Polytechnique de Grenoble University of Exeter Lunds Universitet Université Technologique de Compiègne Uppsala Universitet Institute of Robotik und Mechatronik Hasselt University Faculty of Music, Mc. Gill University of Minnesota • Max-Planck Gesellschaft zur Foerderung der Wiss. E. V. • Université de Genève • Ecole Polytech. Fed. Lausanne • Università Degli Studi Di Padova • Sony France • Ecole des Hautes Etudes en Sciences Sociales • Assoc pour la Création et la Recherche Outils d'Expression • Centre National de la Recherche Scientifique • Université Montpellier-1 Chapel Hill, Oct. 23 rd
CEIT Fraunhofer IIS Erlangen Pisa Stanford U. U. Southern California Chapel Hill, Oct. 23 rd
Labein Queen’s University of Belfast University of Ottawa Labein, Bilbao MIT, Boston Chapel Hill, Oct. 23 rd
Recent activities • P 2 P visualization & NURBS (2000 -02, U. Basque Country) • • • Collisions in mech. Assembly (2001 -05, CEIT) Occlusion culling (2003 -07, CEIT) Improving haptic perception (2006 ---, CEIT) Collaborative haptics (04 -07, Labein) Avatars (04 ---, Vicomtech) Learning Procedural Skills : link IA & VR (07 ---, U. Basque Country - U. Navarre) Chapel Hill, Oct. 23 rd
Let us go to the past : 1986 Chapel Hill, Oct. 23 rd
Outline • • Research environment Reality show or trustable impostor Recent research activities Your input & discussion Chapel Hill, Oct. 23 rd
Reality show or trustable impostor • • • Applications Users, costumers Reality show ? Impostors : interface contract Trustable impostors everywhere Great Expectations – Consumer haptics – Networked haptics • Summing up Chapel Hill, Oct. 23 rd
Applications • Research – Physic based research – Application oriented research • Physics and Maths are the basis of research • Science must also appreciate Applications : – Appraise applied science and engineering ! – Steer knowledge to a good application ! Chapel Hill, Oct. 23 rd
Applications : spreadsheets • David Alan Grier (1979) – “This machine has everything you can find on a real mainframe: multiprocessing, report generators, Cobol. If it were only a small mainframe, there would be no reason to get it, because it is as slow as bad plumbing. However, it has one piece of software. This, my friend, is a spreadsheet and it is going to change the world. Without spreadsheets, these machines are nothing. With spreadsheets, they are unstoppable. ” (source, Computer, sept 2007, pg 6) Chapel Hill, Oct. 23 rd
Applications & users • Windows interface metaphor – Xerox – Apple – Users • Apps target user domains of different size – Specific domains (training, simulation, modeling, … ) – Broad domains (Consumer oriented : games, “desktop” I/F, co-presence, …) • Always – pleased users are the key for successful applications Chapel Hill, Oct. 23 rd
A personal experience : discovering customers • A sad story : ADAMS/Real -Time Kinematics • 1994 : 3 weeks from CEIT to MDI – Jeff Cooper (MDI, Mechanical Dynamics Inc, Ann Arbor MI) • The happy one : ADAMS/Animation – I learnt a new point of view : “customers” (users) – 2 -3 K licenses Chapel Hill, Oct. 23 rd
… (discovering customers) • Both applications (“products”) were : – Easy of use – Quite good “cost / performance” ratio – So, … why this success difference? • ADAMS/Animation – conquered users’ expectations Chapel Hill, Oct. 23 rd
REVIMA (developed at CEIT) • 2000 -04 Collaboration with CEIT – Installed & used successfully – However, it did not replace the space mouse (haptic competitor) desk is more comfortable www. donnarosenartists. com – IEEE Computer Graphics & Applications Dec. 2004 – “Minority Report” Steven Spielberg 2002 film artist Dale Rutter www. ceit. es Chapel Hill, Oct. 23 rd
Reality Show ? • Is the realism of visual or haptic rendering the key for success ? • Who generates the most “realistic” images ? – Sony’s Playstation 3 ? – Microsoft’s Xbox 360 ? – Nintendo´s Wii clearly, doesn’t • So, why did the Wii succeed ? …. . – Why do users enjoy more the Wii ? – The Wii conquered users unconscious wishes – How ? Chapel Hill, Oct. 23 rd
Impostors lead the Wii to success • Does the Wii replicates reality ? • Games for the Wii : Trustable Impostors – An impostor “replaces” something – Replaces a real : golf club, baseball bat, bowling ball, racket – Users trust (accept) their behavior Chapel Hill, Oct. 23 rd
Impostors : replacement restrictions • There is not an absolute replacement – – Impostors can be used to achieve some goals They do not replace playing grounds These are not flaws, they are an Interface Contract Impostors are accepted “as they are”, because they fulfill users' expectations Insertar FOTO juego real Chapel Hill, Oct. 23 rd
Trustable Impostors everywhere • Real time Computer Graphics: – – – Fluid flows Cloth Faked shadows and reflections Herd motion Lighting Vegetation • Simulation – Gravity used to fake acceleration Chapel Hill, Oct. 23 rd
… ( Trustable Impostors everywhere ) • Haptic systems – Deformable environments : body tissue – Comfortable inspection – Pseudo-haptics • Impostors everywhere : – Models “substitute” (real) entities – Algorithms “substitute” physical laws • Mathematical formulations • Heuristics Chapel Hill, Oct. 23 rd
Interface Contract for haptic systems • Impostors are conditioned by hardware specifications, models and algorithms. – Stiffness (mechanical apps) – Force limits (weight, tighten, push, . . . ) – Position precision • Haptic systems require interface contract – Conditions of use (usability domain) analyses must be performed !!! – Perception – Skills achievement Chapel Hill, Oct. 23 rd
90’s Great Expectations : 3 D Visualization • 1991 : Jim Clark (Silicon Graphics) – “He also made some interesting predictions for the low end. He described the multimedia machine of the future as a device integrating the capabilities of (1) full 3 D, interactive, real-time computer graphics, (2) full motion video, (3) image processing capabilities, (4) fully integrated geometric text, and (5) full compact-disc digital sound. He predicts that this will be available in a portable notebook-sized device with motion sensors in the next five to ten years for under $5, 000. ” (Source: Neville Holmes, Computer, sept 2007, pg 11) Chapel Hill, Oct. 23 rd
Great expectations : consumer haptics • A haptic device as spread and affordable as the Wii or the GPU ? – 3 dof seems available – 6 dof and multipoint devices are further away • Which performance ? – Usability domain • Which user expectations will they be able to fulfill ? Which applications should be developed ? Chapel Hill, Oct. 23 rd
Great expectations : networked haptics • Feb. 2007 : Sipping beers with Jarek Rossignac Q-collaborative haptics ? That’s OK. But … which is its use ? A- Mechanical assembly … (Jarek : skeptical glance) A- feel mutual contact … (Jarek : keen glance) • Co-presence awareness – Toast, handshake, – Mutual contact … – Feel physical presence in the collaborative working environment Chapel Hill, Oct. 23 rd
Reality show AND trustable impostors • We must discover unconscious users’ expectations – And fulfill them • “Reality” must be defined by an Interface contract – Users must trust (accept) app. behavior Conditions of use must be well specified Perform more perception and skills achievement analyses • Cheap haptic devices make possible new challenges – May new haptic based metaphors replace the desktop ? – Can haptic devices outperform the mouse ? – Which I/F mechanisms can increase co-presence awareness ? Chapel Hill, Oct. 23 rd
Outline • • Research environment Reality show or trustable impostor Recent research activities Your input & discussion Chapel Hill, Oct. 23 rd
Recent activities • • P 2 P visualization & NURBS (2000 -02, U. Basque C. ) Intell. tutors + 3 D visualization (00 -05, U. Basque C. ) Collisions in mech. Assembly (00 -04, CEIT) (1) Occlusion culling (03 -07, CEIT) (2) Collaborative haptics (04 ---, Labein) Avatars (04 ---, Vicomtech) (3) Improving haptic perception (06 ---, CEIT) IA & VR : Learning Procedural Skills (06 ---, U. Basque C. - U. Navarre) Chapel Hill, Oct. 23 rd
(1) @ CEIT - Simulation group Length 5. 0 feet (1. 5 m) Radius 2. 6 feet (0. 8 m) ITP supplies turbines to Rolls-Royce Chapel Hill, Oct. 23 rd
Multidisciplinar group on haptics 2004 : Comp Graphics Forum Haptic 2007 : IEEE Int. Conf. on Robotics and response Collision + Contact data Automation Other simulations 2008 : IEEE Trans. on Control Systems Technology Control Haptic design & building 2004 : IEEE Comp 2008 : Computer Animation and Virtual Worlds Image generation Graph & APP USER Chapel Hill, Oct. 23 rd
Occlusion culling : environment • Virtual environment – Many objects – Objects not far from eye – Low occlusion • Cannot use – Conventional LOD – “Background texture” replac. Improve occ. cull. Chapel Hill, Oct. 23 rd
Occlusion culling : learnt experiences • Many coherence mechanisms analyzed – Using most of them – Surprisingly stereo (eye 2 eye) coherence did not improve performance … why ? • Low occ. scenes 40% – Speed ups achieved • OC as a LOD mechanism Chapel Hill, Oct. 23 rd
Occlusion culling : learnt experiences • OC as a LOD mechanism – image relevance – Occlusion ratio : • visible pixels / projected pixels – A new GPU query is required – GPU can be easily extended Occlusion Ratio scheme if ( Occlusion_Ratio < threshold & visible pixels < 0. 15% drawing area ) then draw simplified object else draw object normally Chapel Hill, Oct. 23 rd
(2) @ Labein • Collaborative Assembly Simulator • Labein (Spain) & Queen’s University Belfast HAVE’ 2007 , Ottawa, Oct. 13 -14 Chapel Hill, Oct. 23 rd
Learnt experiences • Collaborative haptics (each user holds one object) [ Collaborative ≠ Cooperative two users hold the same object ] – Independent interaction similar to visualization – Dependent interaction • Collisions co-presence • Assembly : constraints, network haptic response – Experiments : Labein - Queen’s U. Belfast – P 2 P & Scalability Chapel Hill, Oct. 23 rd
Current research : motion prediction • Local haptic (not networked) – 5 years ago : 40 ms/frame – coll. computing – Where will be the object when the collision computations are done ? – Free movement vs. contact movement instabilities ? stiffness dependent ? – Currently 2 -3 ms. no relevant • Collaborative haptics (50 -200 ms delays) – promising Chapel Hill, Oct. 23 rd
(3) @ CEIT • Improving haptic perception – Surface discontinuities (edges, holes, thin elements, …) oscillations – Surface normal – Penetration – Contact point Chapel Hill, Oct. 23 rd
Your input & discussion Thanks for your attention Muchas gracias Eskerrik asko (Basque language) Donostia Saint Sebastian Chapel Hill, Oct. 23 rd
b65ac089428cc323983bea5bc8503e46.ppt