1. Contents
Pervasive Computing What is pervasive computing
Current technology
Mobile computing
Context adaptation
Intelligent environment
Morris Sloman
Adaptive architecture
Imperial College, Department of Computing
Security, privacy and management
London UK
m.sloman@doc.ic.ac.uk
www.doc.ic.ac.uk/~mss
4/02/03 Ubiquitous Computing: M. Sloman 2
What is Pervasive Computing? Mobility
Technology View Mobile computing
Computers everywhere – embedded into fridges, washing Computing & communication on the move
machines, door locks, cars, furniture, people
Mostly voice based or embedded?
intelligent environment
Mobile portable computing devices Nomadic computing
Wireless communication – seamless mobile/fixed Intermittent connectivity
User View Usual environment available
Invisible – implicit interaction with your environment Mobile agents
Augmenting human abilities in context of tasks Mobile code and data
Ubiquitous = mobile computing + intelligent
environment
Ubiquitous Computing: M. Sloman 3 Ubiquitous Computing: M. Sloman 4
2. Contents Current Technology
What is pervasive computing
Current technology
Current & near term gadgets
Wearable computing
Mobile computing
Context adaptation
Intelligent environment
Softphone Bluetooth
Adaptive architecture Handsfree
PDA/mobile www.eleksen.com
Security, privacy and management phone
Phone/camera
Ubiquitous Computing: M. Sloman 5 Ubiquitous Computing: M. Sloman 6
Current Technology 2 Wearable Computers
Matchbox Web Server
computer
Best friend
Watch phone Watch camera
Ubiquitous Computing: M. Sloman 7 Ubiquitous Computing: M. Sloman 8
3. Wearable I/O Wearable I/O
LCD
Jacket
Sony Glasstron
Ubiquitous Computing: M. Sloman 9 Ubiquitous Computing: M. Sloman 10
Designer Gear Wearable or luggable?
See http://wearables.www.media.mit.edu/projects/wearables/mithril/index.html
Ubiquitous Computing: M. Sloman 11 Ubiquitous Computing: M. Sloman 12
4. Usability Brainwaves!
Common user interface for workstation and mobile
device applications
Adaptive information display
Replicate characteristics of paper-based notebooks for
annotatablity, robustness, universality
Flexible voice based input-output
Voice recognition + text to speech conversion
Gesture recognition
WAP phone is not a useable computing device!!
Remove human from loop – intelligent agents?
Ubiquitous Computing: M. Sloman 13 Ubiquitous Computing: M. Sloman 14
Contents Mobile Computing Vision
Universal connectivity – anywhere,
What is pervasive computing anytime
Current technology Accommodate heterogeneity of networks
and communicators
Mobile computing
Issues
Ubiquitous intelligent environment –
Wireless communication
embedded computers everywhere
Ad-hoc networking Easy user interaction
Context adaptation Context independent access to services
Intelligent environment + context dependent information
Adaptive architecture
Security, privacy and management
Ubiquitous Computing: M. Sloman 15 Ubiquitous Computing: M. Sloman 16
5. Mobility Issues Wireless Communication
GSM phone 9.6 Kbps,
How to integrate mobile communicators into GPRS typical 28Kbps, theoretical 172Kbps max
complex information infrastructures?
Wireless LAN IEEE 802.11b 200 m range
What effect will they have on work and 2.4 Ghz band: 11 Mbps
leisure? Bluetooth 10 m range
Privacy 2.4 Ghz band: 1 data (700 kbps) & 3 voice channels
How to develop and manage adaptable, UMTS – 3G mobile
context-aware software systems? 114 kbps (vehicle), 384 Kbps (pedestrian),
2 Mbps (stationary)
What support is needed within the network?
HIPERLAN & IEEE 802.11a
Power supplies 5 Ghz band: currently 20 Mbps eventually 54 Mbps
InfraRed – direct line of sight: 4Mbs
Ubiquitous Computing: M. Sloman 17 Ubiquitous Computing: M. Sloman 18
Wireless Problems Ad-hoc networking
Too many similar standards
Shortage of spectrum
Use low power + multiple base stations with
intelligent antenna.
Overlapping spectrum usage can cause interference
eg Bluetooth and IEEE 802.11 Networking with no fixed infrastructure
Use other devices as routers
Unregulated bands lead to chaos
But, security concerns and usage of scarce battery
Health risks? power for relaying – possibly more suited to sensor
than user networks
See http://tonnant.itd.nrl.navy.mil/manet/manet_home.html
Ubiquitous Computing: M. Sloman 19 Ubiquitous Computing: M. Sloman 20
6. Ad-hoc Network Applications Ad-hoc Collaborative Working
Sensor networks Support for cooperative working
Disaster teams Group management protocols
Autonomous robots eg searching buildings, mapping toxic Public and private groups
spills Service discovery
Meetings – exchange visiting cards and information Dynamic adaptation to other device interfaces
Car trains on motorways – 100 KmPH, 2m apart
automatic steering and braking
Ubiquitous Computing: M. Sloman 21 Ubiquitous Computing: M. Sloman 22
Integration of Mobile Systems Contents
Not stand alone devices.
Need to interact with complex legacy information systems eg What is pervasive computing
large databases – merging updates, displaying tables etc. Current technology
Systems development Mobile computing
Requirements specification for adaptable systems Context adaptation
Component composition to meet global QoS, security,
reliability & performance requirements.
Intelligent environment
Mobility models Adaptive architecure
Behaviour specification and analysis Security, privacy and management
Modelling context aware systems
Interaction paradigms
Event-based not object invocation or RPC
Ubiquitous Computing: M. Sloman 23 Ubiquitous Computing: M. Sloman 24
7. Context Awareness Context Adaptation
Context defined by:
Current location Server
Need location detection eg GPS or base station
Indoors – radio beacon, IR
User activity
Walking, driving a car, running for a bus – how to What: Compression, filtering, device-
detect this? specific transformations, information
Ambient environment selection …..
In theatre, alone, in meeting
Local resources or services available Where: Server, proxy or client?
Device capabilities Proxy Proxy client and server do not
Screen, input, processing power, battery life …. change
Current QoS availability – particularly for radio links See Armando Fox work at Stanford
http://swig.stanford.edu/~fox/
Ubiquitous Computing: M. Sloman 25 Ubiquitous Computing: M. Sloman 26
User Selection Map Adaptation
Ultimately, users know best
proxy transforms as best as it can, but gives users a way to
“force” proxy to deliver original content QoS &
here, a simple client-side UI enhancement is coupled with context Rich & dynamic
proxy-side refinement intelligence adaptation, data,
resource Structured data
User context monitoring + metadata
based
selection,
Activity Elements: river, road, motorway, buildings
deadlines Variants: scale, feature detail, date
transformed content local UI refined content Dynamic data: road conditions, weather
from Proxy interaction from Proxy
From presentation by Armando Fox, Stanford
Ubiquitous Computing: M. Sloman 27 Ubiquitous Computing: M. Sloman 28
8. UbiCare: Ubiquitous
Adapting Vector Maps computing for community care
Applications Benefits
Maps can be…
Automated monitoring High lower risk monitoring
Split into features and
presented in part Implanted devices Mobility for chronically ill
Encoded at different scales – Smart clothing Greater out-of-hospital patient
different feature detail Swallow/inject intelligent management
Selective adaptation can sensors and actuators Predictive care from mass data
consider content being Accident and emergency support analysis
degraded Patient record access and
See http://www.doc.ic.ac.uk/~dc/ integration
Wireless video Camera Pill
Ubiquitous Computing: M. Sloman 29 Ubiquitous Computing: M. Sloman 30
Contents Smart Dust
Autonomous sensing and communication in a cubic
What is pervasive computing
millimeter – “dust motes”
Current technology
Sensors for temperature, humidity, light, motion ….
Mobile computing With bidirectional radio or laser + battery
Context adaptation Costs soon < $1
Intelligent environment Typical Applications
Adaptive architecture Defense related battlefield sensors, motion detectors etc.
Security, privacy and management Inventry control on boxes which communicate with crates,
trucks, plane etc to tell you where they are
Product quality monitoring – vibration, humidity, overheating
Car component monitoring
See http://robotics.eecs.berkeley.edu/~pister/SmartDust/
Ubiquitous Computing: M. Sloman 31 Ubiquitous Computing: M. Sloman 32
9. Smart Dust Technology Commercial Motes
Manufactured by Crossbow
http://www.xbow.com
Processor: 4MHz
RAM Memory: 128 Kbytes
On-Board Flash: 512 K bytes
Radio: 916 Mhz, 52K bps
Antenna: On-board, optional external
TinyOS from Berkeley
Ubiquitous Computing: M. Sloman 33 Ubiquitous Computing: M. Sloman 34
Future Smart Dust Pervasive Computer Problems
What means of communication?
Intelligent paper with integrated radio Radio – spectrum shortage
replace current displays Light based – very directional
Smart paint monitors vibrations and detect intruders or Batteries would be impractical power source for 100K
changes colour to react to temperature, lighting etc. processors per person.
Intelligent glass can filter sunlight, become opaque Solar cells are not suitable for all environments.
no need for curtains Solar cells, fuel cells, body heat power?
Smart garments or injectable sensors for people Power not speed is the key issue for future processor
monitoring designs.
Download design and printable motes for < 1c mote SRI Shoe power: 0.5 W
www.media.mit.edu/nanomedia
Printable batteries
http://www.usatoday.com/life/cyber/tech/review/2001-02-12-batteries.htm
Ubiquitous Computing: M. Sloman 35 Ubiquitous Computing: M. Sloman 36
10. Scaling factors Intelligent Environment
> 100K computers per person
Self organising and self configuring
Lights, air conditioning, TV automatically switch on and
Coherent behaviour from vast numbers of unreliable sensors,
off when you enter or leave rooms
actuators and comms. devices Sit on your favourite chair and TV switches on to the
Need new techniques for interaction maybe based on program you usually watch at this time of the day
biological organisms? Use communicator/pda for phone, remote control, keys
Exponential Growth? 94,023 payments, passport, health records, authenticator.
By 2100 ……
billions per Route input from ‘virtual’ keyboard to nearest suitable
mm2 display.
Automatic detection of new items to control and physical
425,352,958,651, 200,000,000,000,000,000,000,000 layout in a room or office, using computer vision.
Billions of computers
Ubiquitous Computing: M. Sloman 37 Ubiquitous Computing: M. Sloman 38
Intelligent Environment? Vision Based Interaction
Presence: Is anyone there? Head tracking
Location: Where are they? Gaze tracking
Identity: Who are they? Lip reading
Activity: What are they Face recognition
doing? Facial expression
Hand tracking
Hand gestures
Arm gestures
From http://research.microsoft.com/easyliving/
Ubiquitous Computing: M. Sloman 39 Ubiquitous Computing: M. Sloman 40
11. Contents Adaptive Application Architecure
What is pervasive computing
Current technology
Mobile computing
Intelligent environment
Adaptive architecture
Security, privacy and management
Remote
Sensors Application
Clients Local & Network
Intermediate Servers Servers
Ubiquitous Computing: M. Sloman 41 Ubiquitous Computing: M. Sloman 42
Policy Policy Based Adaptive Systems
Authorisation policies
Rule governing choices in behaviour of the system Derived from trust relationships to define what
resources or services clients can access, what
Derived from trust relationships, enterprise goals and proxylets or code can be loaded into servers, or
Service level agreements what code loaded into the client can do.
Need to specify and modify policies without coding Obligation Policies
into automated agents Event-condition-action rules to trigger when to
Policies are persistent perform actions, what alarms to generate etc
But can be dynamically modified Ponder – declarative object-oriented language
for specifying policies.
Change system behaviour without modifying See http://www-dse.doc.ic.ac.uk/Research/policies/
implementation – not new functionality
Ubiquitous Computing: M. Sloman 43 Ubiquitous Computing: M. Sloman 44
12. Example Authorisation Policy Obligation Policy
inst auth+ facilities { Defines what actions a subject must do
subject guests; Subject based subject interprets policy and
target gym + pool; performs actions on targets
action enter; Event triggered obligation
when time.between (“0900”, “2100”); Actions can be remote invocations or local scripts
}
Can specify sequencing or concurrency of actions
Ubiquitous Computing: M. Sloman 45 Ubiquitous Computing: M. Sloman 46
Example Obligation Policy Roles
inst oblig heartmonitor { Group of policies with a common subject
subject s = medicAgent; Defines rights (authorisations) and duties (obligation)
on heartanomaly (symptom); Position in organisation – nurse, surgeon
action s.display (messages [symptom]) -> Mobile ‘visitor’ roles in hotel or shopping mall –
send(alarm, symptom); policies which apply to mobile user in an environment
target cardiacCentre;
Paramedic attending an accident
}
Ubiquitous Computing: M. Sloman 47 Ubiquitous Computing: M. Sloman 48
13. Contents Security
Interactions cross multiple organisational boundaries
What is pervasive computing Specification, analysis and integration for
Current technology heterogeneous OS, databases, firewalls, routers
Mobile computing Lessons from history:
Context adaptation Cell phones, IR garage doors, CATV decoders
Intelligent environment Everything worth hacking gets hacked
Adaptive architecture Need for secure ‘out of the box’ set up
Security, privacy and management Identify friend or foe level of trust
Small communicators, with confidential data, are
easily lost or stolen – biometric authentication
Necessary security technology exists
Ubiquitous Computing: M. Sloman 49 Ubiquitous Computing: M. Sloman 50
Privacy Management – the nightmare!
Location service tracks movement to within metres
(cf mobile phones but pay-as-you-go can be anonymous). Huge, complex systems
Clearly indicate you are being sensed or recorded + user control to Billions of processors
stop recording or control distribution of information Multiple organisations
You are now predictable Managing physical world,
System can co-relate location, context and behaviour patterns controlling sensors, actuators
Do you want employer, colleagues or insurance company to know you Humans will be in the way
carry a medical monitor? Errors propagate to bring down complete regions
Tension between authentication and anonymity – business want to Hacker and virus paradise
authenticate you for financial transactions and to provide ‘personalized’
service cf web sites System propagates false information about
Constant spam of context dependent advertising
individuals or organisation
Complexity of s/w installation on a workstation or
server – how do you cope with billions?
Ubiquitous Computing: M. Sloman 51 Ubiquitous Computing: M. Sloman 52
14. Management Solutions
Policy
Intelligent agents, mobile agents, policy
QoS Management
Fat pipes and large storage can convert media streams to short
traffic bursts in core network but still needed for wireless links
Adaptive self-management is the only answer
Partitioned domains of responsibility
Genetic algorithms may be suitable for long-term strategy but need
more deterministic solutions for short term decision making
Remove human from the loop
Ubiquitous Computing: M. Sloman 53 Ubiquitous Computing: M. Sloman 54
Conclusions References
Pervasive Computing
IEEE Pervasive Computing Magazine
Universal PDA/communicator Intelligent Environment
Explosion in embedded sensors/actuators http://www.media.mit.edu/
http://cooltown.hp.com/
Context-aware intelligent environment http://portolano.cs.washington.edu/
Privacy will be a major issue http://www.firstmonday.dk/issues/issue4_9/odlyzko/
Out of the box security Wearable Computers
http://www.redwoodhouse.com/wearable/
Adaptive self-management is needed – biological http://iswc.gatech.edu/archives.htm
paradigms? Wireless communications
http://www.wirelessdevnet.com/
Mobile computing
http://computer.org/dsonline/
http://www.mobileinfo.com
Ubiquitous Computing: M. Sloman 55
http://www.comp.lancs.ac.uk/computing/research/mpg/most/
Ubiquitous Computing: M. Sloman 56