490 likes | 600 Views
Pervasive Computing for the 21 st Centuary. Agents, Virtuality & Mobility. Gregory O’Hare Department of Computer Science University College Dublin. University College Dublin DEPARTMENT OF COMPUTER SCIENCE COMP 4.19Multi-Agent Systems(MAS) Lectures 15-16.
E N D
Pervasive Computing for the 21st Centuary Agents, Virtuality & Mobility Gregory O’Hare Department of Computer Science University College Dublin
University College Dublin DEPARTMENT OF COMPUTER SCIENCE COMP 4.19Multi-Agent Systems(MAS) Lectures 15-16
Some Thoughts : A Difficult Journey • The Volume of Data and the expectations in terms of ability to process and • asimulate this data is ever increasing; • Content will become all pervasive; • Traditional Inforamtion Space Boundaries will begin to decay;
A Necessary Synergy Agents Mobility Virtuality
Another Interesting Synergy Embodiment Situatedness Immersion
Mobile & Ubiquitious Computing Computing as a discipline has and is undergoing a journey. Initially computing was viewed as a desktop activity, the desktop metaphor then…. The laptop metaphor…………now the palmtop metaphor as exemplified by the emergence of the Personal Digital Assistant (PDA). The ubiquitious computing metaphor is almost upon us where devices are computationally empowered and can in a non intrusive manner communicate and collaborate (eg Bluetooth Enabled Devices). Wearable and Digestable computing awaits us. In short all pervasive computing.
We regard mobile computing as involving the use of intelligent agents • that are truly mobile in nature. • We characterise mobility across several dimensions :- • Semi-autonomous entities that roam a physical environment • ( like robots ); • 2. Entities propelled through a physical space by a user, • like handheld advisory systems located on palm top • computing devices; • 3. Avatars immersed & Navigating within a Virtual World; Mobile Computing
The Information Explosion Today Information is power. The Internet represents the most extensive global information resource. The web structure is characterised as a graph like structure comprising of over 1 Billion nodes (individual pages) and billions of connecting edges (hyperlinks). A recent survey of 200 Million pages had 1.5 Billion associated hyperlinks (Henzinger et al 2000). It is anticipated that the web graph is 1 Billion nodes and expanding at a rate of 1 million per day ………….The Information Explosion
Personalisation & Contextualisation The key to ameliorating the Information explosion is Personalisation and Contextualisation of Information Content. In Essence we wish to supply the mobile user with content when they need it where they need it and in a form relevant to their technological context. Proactive (Push) tailored content delivery.
Hyper Interaction in Physical Space • HIPS (Hyper Interaction in Physical Space) is an initiative to facilitate the intelligent dissemination of information to tourists • Esprit long term research project supported under the I3 net (Intelligent Information Interfaces) initiative - Project# 25574 • (O’Grady & O’Hare 1999 & O’Rafferty & O’Hare 1999)
HIPS Objectives • Provide an infrastructure which supports simultaneous user navigation of a physical space and it’s corresponding information space • Deliver contextualised information to roaming tourists • Provide seamless movement between the physical & virtual spaces
Localisation Technologies • A variety of Approaches exist:- • Satelite based Systems • Global Positioning System (GPS); • Differential GPS (DGPS); • Galilleo - European • GLONASS – Russion; • Radio Based Systems • Base Cell Identification; • Base Cell Triangulation based on either Angle of Arrival (AOA) or Time of Arrival (TOA); • Location Fingerprinting – each location has a unique signal pattern create database to record these and match; • Hybrid Systems; • Enhanced/Assisted GPS (EGPS) – Ericsson Mobile Positioning System (MPS).
HIPS : A Handheld Tourist Aid Tourist Content Delivered in a Proactive (Push) Tailored fashion. The key personalisation & Contextualisation parameters are: Location (Longitude, Latitude, Orientation) acquired from a Global Positioning System (GPS) device & electronic compass; User Preferences extracted from a user profile;
AD-Me System • AD-Me (Advertising Delivered to the Mobile E-commerce user) System • (Hristova & O’Hare 2001) Seeks to :- • Offer context sensitive advertising; • Across a miriad of mobile devices; • Accomodating a rich variety of formats protocols like HTML, WMl, imode; • Embraces an Multi-Agent Systems (MAS) approach to deliver proactive • push system functationality;
Where Are You? • “I’m walking down Grafton St.” • Most Common Use of Mobiles. • Allow the Mobiles Themselves To Deal with these Requests. • Context Aware Computing.
Objectives • Support the mobile citizen in the location, tracking and rendezvousing with a variety of moving entities. • Develop an lightweight mobile agent based infrastructure to support mobile users. • Utilise GPS in the location of individual entities. • Create a versatile, user friendly GUI. How ?
WAY GUI Zoomed In Zoomed Out
Context Sensitive Systems • CyberGuide • Mobile context aware tour guide, based on tourist position and orientation. • Makes use of indoor and outdoor location positioning systems. • HIPS • Innovative system delivering personalized multimedia information specific to users location. • Prototype handheld touist guide. • Others • Active Badge and ParcTab.
Agent Based Context Sensitive Systems • comMotion • Uses GPS and a behaviour learning agent to learn a persons daily travel routine. • Provides location based information e.g. nearest bank. • Uses GUI to display maps and control administrative functions. • Impulse • Uses agents to query a database and provide a limited range of relevant resposes to the user. • Others • Ad-me and DealFinder.
WAY Components • Personal Digital Assistant (IPAQ 3600 - 3700). • Pen-based interface. • Nokia Card Phone 2 GSM 900/1800. • GPS card / Garmin 2 GPS receiver. • PCM/CIA sleeve. • Agent Factory Server.
Existing Mobile Agent Systems • JADE (Telecom Italia Lab 1999) • LEAP (European Project 2000) • JAM (Huber 1997) • Ara (Peine et al.1997) • D’Agents (Gray et al. 1999) • Grasshopper (Baumer et al. 1999)
WAY Agents • BDI architecture • Belief Desire Intention • Mobile • Can migrate from platform to platform • Lightweight • Minimises software footprint • Eases migration load • Java • Portability • Availability on a variety of platforms
Agent Factory LITE • Java implementation of AF run-time environment. • Can interpret and interact with Smalltalk AF agents. • Minimises the software footprint on PDAs.
PDA WAY agent mental state + Address Book subscribe? Agent Factory Server
WAY Connection Request Location? Request WAY Connection? Refuse WAY Connection WAY Connection Inform Location
Can’t Connect Low Power or Shutdown Agent migrates to server Request Last Known Position? Inform Last Known Position
Agent Mental State: Time 1 Beliefs Time: 1 NEXT(BELIEF(checkpower)) ALWAYS(BELIEF(name(Bill))) BELIEF(location(Self,53.34436,-6.26129,NW)) BELIEF(requested(position,Joe)) BELIEF(wayCon(Joe)) BELIEF(wayCon(Sue)) BELIEF(wayCon(Sue)) BELIEF(way(Sue)) BELIEF(way(Sue)) Commitment Rules BELIEF(location(Self,?long,?lat,?ori)) & BELIEF(requested(location,?user)) & BELIEF(wayCon(?user)) => COMMIT(Self,Now,inform(?user,location(Self,?long,?lat,?ori))) BELIEF(way(?user)) & BELIEF(wayCon(?user)) => COMMIT(Self,Now,request(?user,location))) Commitments Time: 1 COMMIT(Self,Now,inform(Joe,location(Self,53.34436, -6.26129,NW))) COMMIT(Self,Now,request(Sue,location))
Agent Mental State: Time 2 Beliefs Time: 2 ALWAYS(BELIEF(name(Bill))) BELIEF(checkpower) BELIEF(wayCon(Joe)) BELIEF(wayCon(Sue)) BELIEF(informed(location(Self,53.49835,-6.25777,S),Sue)) Commitment Rules BELIEF(informed(location(Self,?long,?lat,?ori),?user)) => COMMIT(Self,Now,updateMap(?user,?long,?lat,?ori)) Commitments Time: 2 COMMIT(Self,Now,updateMap(Sue,53.49835,-6.25777,S))
AF Migration • Weak migration • Smalltalk/Java interaction • Agent design including mental state • Less traffic over wireless network
Humanoid Robots Honda P-3 Robot Sony Aibo
Social Robot Technologies • RoboCup • Fundamental Research • Fukuda Lab (Robotics) • Uni. of Reading / Edinburgh: K. Dautenhahn, A. Billard (Alife) • Manchester Metropolitan Uni, CPM: Bruce Edmonds (Mgmt & business - Socially Situated Intelligence, Embeddedness) • University of Oregon: Social Interaction Lab (Psychology) • University College Dublin Social Robotic Architecture (SRA)
Reality & Virtual Reality Real World Virtual World
Collaborative Virtual Environments • Collaborative Virtual Environments (CVE’s) represent a • broad class of environment within which………….. • Users are immersed within a Virtual World through • personalisable avatars; • Users are furthermore immersed within the associated • Virtual Community; • The CVE must provide a Personalised and Contextualised user • experience. • Again user tracking is crucial in order to compile & maintain user profiles. • Within Virtual Reality Modelling Language (VRML) worlds this is • achieved via the judicious embedding of sensors (touch, proximity, time). • The ECHOES System is one such system (O’hare et al 2000)
Virtual World Personalisation Personalisation of these worlds can be achieved through the dynamic reconfiguration of the world to reflect the individual user preferences. The ENTER System is one such system and was pioneering in this respect; (Guinan, O’hare & Doikov 2000) Thus in an ecommerc context relevant retail outlets are pulled to the foreground whilst less relevant outlets are pushed to the background. Relatively simple user profiling can yield significant personalisation. Thus we can deliver powerful personalised market penetration that cannot be achieved within its physical world counterpart.