CS 290-63: Ubiquitous Computing

Prasun Dewan

SN 150 Department of Computer Science

University of North Carolina dewan@unc.edu

www.cs.unc.edu/ubiq

TR: ???Office Hours: ???

(Prsün Divän)

2

everywhere indefinite number of times

Ubiquitous Computing

Ubiquitous

Computing CPU, Memory, Communication

Not really possible or desirable more than in traditional

computing so widespread that

unnoticed

3

Traditional Computing

Desktop Computers /Servers

Embedded Computers

4

Ubiquitous = More of the Same?

No new technical issue Except scalable

networking IPV6

In ubiquitous computing we see new kinds of computing

5

Interactive Mobile Computers Motivation

Pervasive access to standalone apps Calculator, Games, Music, Video

Issues Hardware realization (will be

mostly ignored) more robust lower-power lighter weight

6

Interactive Mobile Software Issues

Resource-poor power

• Scheduling, result accuracy

I/O devices • Speech, Handwriting recognition

Storage• MEMS

Tiny OS Security

More vulnerable to theft and destruction

Exposed to more people

7

Strongly-Connected Mobile Motivation

Remote data access web, rlogin

Collaboration Email, chat, application-sharing

Computation offloadingInternet

8

Strongly-Connected Mobile Issues

Mobile IP Address contains location but

location variable Adapting to mix of heavyweight,

lightweight computers Dynamic application-sharing

architectures Multi-device collaboration

Internet

9

Intermittent Connection Disconnected and strongly-

connected states Remote data accessed in

disconnected state

Motivation Work in the offline mode

Paper, spreadsheets

Issues Local database smaller

hoarding Conflicts can arise

merging

Internet

10

Wireless (Weak) Connection Additional weakly connected

state higher delay, error rate

Motivation Ubiquitous remote access

Network

11

Wireless (Weak) Connection Issues

Physical Layer Bluetooth, IEEE 802.11, Infrared

Optimizing weak connections Batching changes propagated for

merging Transcoding

Adapting to multiple connection levels Adaptive TCP/IP Adaptive Transcoding Adaptive update propagation for

merging

Network

12

Networked Embedded Computers

Network

networked appliances sensors historical sites & other locations

Connected to network send and/or receive

May be embedded only for network access

13

Peer of Networked Embedded Computer?

Network

Another embedded computer

Non embedded Computer

14

extendibility & scalability

Embedded Peer

Network

Composite devices security system

Distributed composites vs. hardwired devices

client-defined composites

ease of change

reuse of constituents

15

“Smart Dust”

Network

Late binding Monitoring

Temperatures in a Fire

Monitoring rain levels in a flood

16

Embedded Computer communicating with Non embedded Computer

Network

Embedded is client Embedded is server

17

Embedded is Client

Network

viewing/recording status image from camera body part data instrument visualization

Motivation

trigger orders refrigerator items

18

Embedded is Server

Network

Interactive computer issues commands to embedded computer

19

Motivation

A single remote control for arbitrary (dissimilar) appliances TVs, EKG machines, Water

Sprinklers, Car Seats

Interactive computer can offer different user interfaces from conventional controls

Automatic appliance- remote control binding

Control appliance from arbitrary locations Water sprinkler from

beachInternet

20

Per-User Control of Shared Appliances

Mobile device knows owner’s info: Favorite channels Volume preferences Credit card PIN Files to be printed Car-seat tilt angle

Record/replay Pin recorded first time

and replayed later

21

More Powerful Device Output Device

EKG results Refrigerator contents Sprinkler settings

More computation power Undo/Redo Cut/Paste Save Macros

Enter arbitrary input Water sprinkler days

Offline Mode Synchronization

22

Control Appliance Collections

Multicast commands Start/stop recording

images from all cameras

Switch off all lights in the room

Switch of all appliances except the refrigerator

23

Control Appliance Collections

Network

Distributed presentation On projector, speakers,

computer screens

24

Motivation Summary More universal

More control

More remote

Internet

25

Networked Embedded Computers

Network

Issues Late binding

Naming Discovery IPC User-interface deployment Multi-appliance control

Access control Existing social protocols not

supported by existing mechanisms• All co-located users can use

appliance

• Children can see certain channels

26

Ad-Hoc Networks Peer to peer connections

among late bound computers

Motivation Spontaneous collaboration

classroom, war, airport meeting Sensor networks

disaster recovery• e.g. heat sensors thrown on fire

Network

27

Ad-Hoc Networks Issues

No Router ad-hoc routing

• Power-aware

Network

28

Location-aware computing operations based on current,

past, and future locations of users and devices

29

Location-Aware Computing Motivation

location-based action nearby local printer, doctor nearby remote phone directions/maps

location-based information real

• person’s location• history/sales/events

virtual• walkthrough• story of city

augmented• touring machine

30

Pose-Aware Computing Operations based on

locations and orientations of users and devices

Motivation Augmented reality

Magic Lens

31

Wearable Pose-Aware Computers Computers on body

track body relative movements monitor person train person

32

Location/Pose- Aware Computing

Issues Tracking algorithms

fine or coarse grained Software architecture Maintaining privacy

33

Summary Traditional

Stationary desktops/servers and Embedded

Ubiquitous Interactive mobile

resource-poor security vulnerability

Strongly-connected mobile mobile IP adapting to heavyweight/lightweight mix

34

Summary Ubiquitous computing

Intermittently connected mobile hoarding synchronization

Wireless (weak) communication Physical layer Optimizing weak connections Adaptations for multiple connection levels

35

Summary Ubiquitous computing

Networked embedded computers Naming Discovery IPC UI deployment Access control

Ad-hoc networking with late bound devices ad-hoc routing

Location and Pose aware Tracking Architecture Privacy

Wearable Avoid strong signals Displays

36

EmbeddedMobile

Interactive

Beyond Desktops/Servers

Wearable Active badge

Location

SensorFlight

Simulator