Grounding to the Real World - Architecture for Ubiquitous...

Preview:

Citation preview

Grounding to the Real WorldGrounding to the Real World-- Architecture for Ubiquitous Architecture for Ubiquitous

Computing Computing --ISMIS 2003ISMIS 2003

Hideyuki NakashimaHideyuki Nakashima中島秀之中島秀之

Cyber Assist Research Center, AISTCyber Assist Research Center, AIST

h.nakashima@aist.go.jp 2

Background

• Information technologies are advanced to the point where it becomes feasible to assist ordinary people enjoy the full capability of information processing powers.– Although the Internet opened up huge

possibilities for daily use, users still have to learn a lot of special concept to use it.

h.nakashima@aist.go.jp 3

Background 2

• With the help of various sensing technologies, it is time to design arich, keyboard-less interface and provide an ubiquitous computing environment for ordinary people in everyday life. Ubiquitous agent technology plays essential role here.

IntroductionIntroduction

h.nakashima@aist.go.jp 5

TICKET

Take the next train.Change to A line at Ueno station.The fare is YY yen.

I want to go to Maebashi station.

Good Old DaysGood Old Days

h.nakashima@aist.go.jp 6

券売機券売機 券売機

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

券売機

・・・・・

When she getsolder...

When she getsolder...

h.nakashima@aist.go.jp 7

券売機 券売機 券売機 券売機

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

How much does it cost?

What is the best route to Maebashistation?

h.nakashima@aist.go.jp 8

券売機 券売機 券売機 券売機

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

●●●●

××××

△△△

○○○○○

▲▲▲▲

120

560

340

780

900

□□□ 210

※※※※ 400

※●※●※

×○×

△■△■

○□○□

▲×▲×▲

120

560

340

780

900

□◎□◎ 210

■※■ 400

あっ!

h.nakashima@aist.go.jp 9

Please go through..

The next train will leave from platformNo 2 to your right.

TURN RIGHT TURN RIGHT

speaker

Connection to Internet

Time table server

Seat reservation system

My-button

LED display

Station master computer

Station LAN

TokyoRF tag

Unawareable environmentUnawareable environmentIR communicator

Position sensor

h.nakashima@aist.go.jp 10

Ubiquitous Computing• Omnipresence of computing power

• Human centered services– The opposite direction from virtual reality where the

reality is reconstructed in the digital world

– Computers come out to the real world

• Gadgets are hidden– calm computing– disappearing computer

– oxygen

– unawarable computers

h.nakashima@aist.go.jp 11

MIT Oxygen: Intelligent Space• Context aware environment: computation “in the air”

Stereo Cameras: to track persons and objects

Stereo Cameras: to track persons and objects

Active Cameras: for video-conferencing and meeting archive support

Active Cameras: for video-conferencing and meeting archive support

Microphone Arrays: beam-forming for untethered spoken interface

Microphone Arrays: beam-forming for untethered spoken interface

h.nakashima@aist.go.jp 12

Before:People had to learn to use machines

Computers in the Digital WorldComputers in the Digital World

People in the Real WorldPeople in the Real World

Internet schoolInternet schoolPC schoolPC school

h.nakashima@aist.go.jp 13

After:Grounding is the key technology

Digital WorldDigital World

Real WorldReal World

groundinggrounding

h.nakashima@aist.go.jp 14

“Grounding”

• Connection between digital data and real things– similar to “symbol grounding problem”

– (but not the same)

• Total grounding is not necessary nor possible– utilization of artificial means or human aid

h.nakashima@aist.go.jp 15

From WWW to HNM

• Wherever, Whenever, Whoever

→ Here, Now, Me (situatedness)

• At the same time:

Just as much as you want to know

→ intelligent information filter

• And,

Just as much as you want to be known

→ privacy

Location based communicationLocation based communication

h.nakashima@aist.go.jp 17

Human Centered Infrastructure

ID-based communication

ID-based communication

Location-based communication

Location-based communication

Broadcasting satellite36000km~780km

Broadcasting satellite36000km~780km

High-altitude aeroplane/ship20km

High-altitude aeroplane/ship20km

broadcastbroadcast

broadcastbroadcast

Sensor networkSensor network

h.nakashima@aist.go.jp 18

Communication Methods

Local 1-10 meters global

Magnetic cardMagnetic cardICIC--cardcard

NetworksNetworksPhonesPhones

BroadbandHigh-speed

LowLow--powerpowerShortShort--rangerange

LocationLocation--basedbased

IDID--basedbased

h.nakashima@aist.go.jp 19

Advantages of Location Based Communication

<x, y, z, t>

• Location is one of the essential keys for grounding the digital (logical) world to the real world.– Situated information processing and services

• Interaction without ID– Possibility of PRIVACY protection

– For better human interface

– Application to emergency situations

h.nakashima@aist.go.jp 20

Various location-based communication(addressing) methods

(cricket)*(mobile phones)*

(cricket)*(mobile phones)* carrier pigeoncarrier pigeon

Logical addressing“passenger at 12C”

Positioning by server

Self-Positioning

Physical addressing“you”

(bluetooth)*DSRC

(bluetooth)*DSRC CoBITCoBIT

*positioning only

h.nakashima@aist.go.jp 21

Security vs. Privacy vs. Convenience

securitysecurity privacyprivacy

convenienceconvenienceThey basically contradict each other with some exceptions

h.nakashima@aist.go.jp 22

Needs technology for privacy protection

amountofservice

privacy protection

not zero

h.nakashima@aist.go.jp 23

Dynamic/Voluntary Control of the Balancing Point is Necessary

• Trade-off between the amount of services and the amount of personal information

• Let the user set the point!

Privacy Security

AgentAgent--based Network based Network ArchitectureArchitecture

h.nakashima@aist.go.jp 25

Ubiquitous Computing

How do we access and control devices

Intuitive User Interfacethat uses real world propertiessuch as location(point and click)

h.nakashima@aist.go.jp 26

Why Agents?

• Aids for human cognition– interface

• Resource control– cognitive resources

– devices• multiple users and• multiple devices

h.nakashima@aist.go.jp 27

Cognitive Resource Control

• SmartKom example:– cognitive resource of human

• when running terse• when walking• when standing verbose

h.nakashima@aist.go.jp 28

DFKI: SmartKom

• positioning– indoor: light beacon

– outdoor: GPS

h.nakashima@aist.go.jp 29

・Maintains who reserved which time slot・Controls remote access・The reserver has the access right

homeVCR

toilet

bathliving

air conditionerthermometer

thermometer

・The access right belongs to who reserved・Access right can be delegated・The device may act as a mediator

homeserver

RFID tagRFID tab

RFID tag

Son

×

×

town

OK Dad, go ahead...

Access Control of Home AppliancesBased on Location and Priority

My son has a reservation. Let’s ask him.

someone is using the air conditioner・change reservation

or, delegate authority

・Deny remote access when someone is near-by

・communication center for the family

h.nakashima@aist.go.jp 30

CONSORTS

InferenceEngines

Personal Agents

ServiceAgentsUser1 Usern

Communication and censoring

devices

Physical WorldContentServer

Devicewrapper

Agent communicationAPIphysical link

Spatial information database- sensor-based geometrical representation- ontology- location information

Spatio-Temporal Reasoning Agent

Camera, Wireless LANCoBIT, etc.

User modelmanager

Physical Grounding

Web Services

h.nakashima@aist.go.jp 31

Spatio-temporal Inference Agent in Consorts

spatio-temporal map(Segment Representation)

Inference engine:spatio-temporal relations and their changes (distance, phase, connection etc.)

device wrapper agents

personal agents,service agents

request, reply, inform

time

sensor data

Spatial informationdatabase

h.nakashima@aist.go.jp 32

interpretation

Sensor information at location (x,y) mapped onto a map

Geometric representation of sensor information

Museum

1st Floor

Entrance

3rd Floor2nd Floor

Room CRoom A Room B

Painting BPainting A Painting C

Hierarchical representation of physical space

Painting D Painting D

h.nakashima@aist.go.jp 33

Ubiquitous Computing with FIPA-Agents

• Direct connections to sensors are not good

• Agents mediate communication of sensor information

• Agentcities Network• http://www.agentcities.net/

– EU centered worldwide research network based on FIPA standard

– 134 platforms• EPFL, BT, Motorola, etc.

– Tokyo.Agentcities.Net (CARC)• http://www.carc.aist.go.jp/agentcities/

h.nakashima@aist.go.jp 34

A Snapshot of Our Museum Information Support System

Main Window (Museum Map)

Information Window

Message WindowMessage Window

Human Icon: Current Location

Yellow Area: Service Zones

Blue Line: Users’ Trajectories

h.nakashima@aist.go.jp 35

UBKit

UBKit: Software construction tool kits

middleware

infrared

SOAP DNS

network

IEEE802.11 BluetoothMicroservers

voice positioning

real world

CONSORTSCONSORTS

h.nakashima@aist.go.jp 36

Micro servers + Optical Communication forIntegrated Remote Control System

Ultra small low energy micro server(Linux) coin size

Self configuring plug and play network

radio

positioning system

voice recognition

sensor networkinfrared

Lightweight middleware (service directory, SOAP)

sensorgateway

UBKit (Ubiquity Building Toolkit)

Integrated control of all home applianceswith intuitive interface

Connection toConnection tothe Semantic Webthe Semantic Web

h.nakashima@aist.go.jp 38

Annotation as Grounding

Content(Real World)

SemanticWeb

AnnotationAnnotation

Semantics forComputers

Sharing semanticsbetween computers

and human(Grounding)

h.nakashima@aist.go.jp 39

Multimodal Content Processing

• Grounding linguistic content to the physical world– annotation with location– interaction through CoBIT

• User modeling– record of movement– record of inquiries

• Summarization and interactive presentation

Cyber Assist ProjectCyber Assist Project

h.nakashima@aist.go.jp 41

The goal:Grounding the digital world to the real worldfor “here, now, and to me” service

informationinformation

Digital worldDigital world

Real world

goodsgoods

humanhumanSemantic WebSemantic Web

Ubiquitous ComputingUbiquitous Computing

h.nakashima@aist.go.jp 42

My Button• One button

– Do what I want here, now.

– No button?Activation by

brain wave?

• The ultimate man-machine interface?– Intelligent

personal agent– Sharing

situations

command button

fingerprint recognizer

reflectors photo detector

microphone

speaker

h.nakashima@aist.go.jp 43

Optical Bi-directional Communication

server

cameraLED

terminal

•Download•voice

•power supply

•Upload•position•direction•motion•(ID)

h.nakashima@aist.go.jp 44

My Button V1: CoBIT

• Compact Battery-less Information Terminal

• Terminal: solar cell + reflector + earphone

h.nakashima@aist.go.jp 45

My Button V1

• Server (environment) side: LED + camera(s)

h.nakashima@aist.go.jp 46

青森産のリンゴです。一個200円です。これ買いますか?

ありがとうございます。

アラスカ産のシャケです。

これを買いますか?

となりにはサンマがあります。

Application to Markets

h.nakashima@aist.go.jp 47

Environment

The Image of the Target System

User

device

Intelligent Content

Sensors and actuators

Situationrecognition

Query/orders

information

request

Information retrievalInformation processing

Voicerecognition

Semantic query

User modeling

Photo/radio communication

vision

Multiagent technology

Application toEmergency rescue

h.nakashima@aist.go.jp 48

Tokyo

WirelessLAN

My-Button

i-lidar

camera

intelligent contentsemantic web

voicecommand/query

voiceinformation

IR

The next train will leave from platformNo 2 to your right.

CoBIT

h.nakashima@aist.go.jp 49

The evolution of the communication technologies in Star Trek

CoBITmy buttons

conventional

ApplicationsApplications

h.nakashima@aist.go.jp 51

Mass Support

• Information support for the whole city

• Examples– Global Car Navigation

• Optimal route planning using the current positions and destinations of all cars in the area

• Coordinated signal control

– Theme Park total coordination

– On-line Government• Direct democracy with the help of discussion and

decision support system

h.nakashima@aist.go.jp 52

Special project for earthquake disaster mitigation in urban areas(Ministry of Education and Science)

• Ad hoc network for emergency communications

• Multiagent simulation for– evacuation

– rescue

h.nakashima@aist.go.jp 53

Real-time dynamic ad-hoc network for large-scale disasters (earthquake)– Effective in

congested areas

– Alternative to other communication infrastructures• First 10-30 min.

– Connection to other lines when possible

h.nakashima@aist.go.jp 54

Grounding the Simulation

KernelGIS

Agents

Viewers

ComponentSimulators

Portable devices Robots

Rescue team

Real-worldsensors

Rescue HQ

h.nakashima@aist.go.jp 55

Summary

• We have to connect the digital world to the real world→ GROUNDING

• Multiagent concept is the key to ubiquitous computing– Grounding

– The core architecture

– Ubiquitous agents for human support– Many MA applications