Enabling Technologies for Security in Ambient Intelligence

1
Enabling Technologies forSecurity in Ambient
Intelligence

• Pieter Hartel

(Read N. J. Henderson, P. H. Hartel, Pressure
Sequence, in 4th Int. IFIP wg 8.8 Conf. Smart
card research and advanced application, Kluwer
Academic Publishers, Sep. 2000, pp 241-256 L. R.
Taylor, Making Guns Safer, National Institute of
Justice Journal, Jun. 2000, pp.16-19)
2
Overview

• Ami definition and requirements
• Enabling technologies
• Two case studies
• Conclusion

3
Definition

• Ubiquitous computing
• Ubiquitous communication
• Intelligent user-friendly interfaces

• (source IST Advisory Group
• Scenarios for Ambient intelligence in 2010
• European Commission, Jul. 2001)

4
Requirements

• Unobtrusive hardware
• Seamless mobile/fixed communications
  infrastructure
• Dynamic and massively distributed device networks
• Natural feeling human interfaces
• Dependability and security

5
Enabling Technologies forSecurity

• Micro payment
• Smart cards
• Verification
• Biometrics

6
Case Studies Smart Objects

• Can they provide feedback on the way the object
  is held?
• Can objects authenticate the rightful owner/user?

7
US Statistics

• 500,000 guns stolen each year
• 7 Law Enforcement officers killed each year with
  their own weapon
• in 1999, 28.874 died from firearm injuries, 11.7
  were under 20

8
Smart Gun
9
Requirements

• Operate reliably
• Have all capabilities of a fire arm
• Able to be fired by other officers
• Easy to operate and maintain
• Verify and approve user in time
• Include indicator when enabled
• Fire even if the electronics fail

10
Fingertip Position
11
Results

• 160 subjects
• Promising results but 20 times better needed
• Future work include pressure info

12
Statistics

• World wide 1 Billion smart cards sold per annum
• 0.5 Million mobiles stolen/lost p.a. in the UK
  alone

13
Biometrics Associates
14
Pressure Sensor
15
Piezo Electric Pressure Sensor

• Applied pressure causes proportional charge
  generation

16
Experiment

• 34 volunteers
• Each chose their own sequence
• Sequences between 2 and 14 taps

• 8 enrolment
• 8 test sequences
• gt1000 impostor sequences

17
Signals
18
Receiver Operating Characteristics

• 2.3 equal error rate!

19
Results

• Simple (Tap a rhythm)
• Fast (lt 0.2s in Java on iButton)
• Reliable (EER of 2.3 with 34 users)
• Inexpensive (2 cent/sensor mass prod.)

20
System Issues

• Biometric templates are private!
• Secure enrollment
• Secure storage of template
• Hardware tamper resistance
• Theft of devices is a problem
• 0.5 Million Mobiles stolen in UK in 2001

21
Conclusions

• Objects can be made smart
• The context is important
• Engineering challenges
• Business opportunities
• Security is only an aspect of AmI

22
Handout
23
Classify your fingerprints
www.poof-slinky.com