How to Build Smart Appliances

1
How to Build Smart Appliances?

• Babak Youssefzadeh
• Computer Science Engineering
• University of Texas at Arlington

2
Smart Appliances?

• Smart Appliances are characterized as devices
  that are attentive to their environment.
• They are devices that are not ignorant about
  their environment and context of their usage.

3
Topics Covered

• Context
• Ideas for capturing information to approximate
  context
• Mechanisms and Methods
• Building Applications

4
Context-Aware Devices

• In the 90s, context-aware devices were
  synonymous to location-aware devices.
• In the Technology for Enabled Awareness (TEA)
  project, Context-awareness is defined as
  Knowledge about the users and devices state
  including surroundings, situation and location.

5
Real Situations
6
Capturing Data

• Type of Sensors
• Light sensors
• Cameras
• Microphones
• Temperature
• Air pressure
• Location
• Motion Detectors
• Accelerometers

7
Capturing Data (cont)

• Accelerometers
• Inclination
• Motion
• Acceleration
• Examples
• Mercury switches
• Angular sensors
• Accelerometers

8
Capturing Data (cont)

• Location Sensors
• Position, proximity location and collocation of
  users, devices and environment.
• Examples
• GPS
• Active Badge

9
Capturing Data (cont)

• Touch Sensors
• Used mainly when the devices are to be operative
  in the users hand
• Skin conductance directly
• Light or temperature sensors indirectly

10
Capturing Data (cont)

• Biosensors
• Measure skin resistance, blood pressure etc.
• Mainly used for Sporting and Medical applications
• emotional state of the user may also be obtained

11
Capturing Data (cont)

• No Power sensors
• Designed for extremely low power consumption
• Controlling Micro controllers
• Examples
• Metal ball switches
• Mercury Switches
• Solar panels

12
Constraints on Data Capturing

• Portability, usability and design
• Power consumption
• Calibration
• Setup Time
• Reliability
• Price and Cost
• Unobtrusiveness

13
Architecture from sensor data to Applications
14
Cues

• Provide an abstraction from physical and logical
  sensors
• Each cue is dependent on one sensor but a sensor
  can provide many cues.
• Only cues need to be adapted to include new
  sensors with different characteristics.

15
Cues (cont)

• Examples
• Average
• Standard deviation
• Quartile distance
• Base frequency
• First derivative
• Cues are an useful abstraction for communication
  in distributed sensing devices.

16
Real Example

• A context is a description of the current
  situation on an abstract level.
• Example
• If ( light is low and acceleration is nonzero )
• Then context Moving in the dark

17
Distributing Computation Load

• Kohonens self-organizing maps
• Use clustering algorithms
• Learn clusters dynamically
• Clusters and context can be used to predict later
  contexts
• http//www.mlab.uiah.fi/timo/som/thesis-som.html

18
Steps to build Aware devices

• Step 1 Identify the contexts that matter
• Step 2 Find the appropriate sensor
• Step 3 Build and assess a prototypical sensing
  device
• Step 4 Determine recognition and Abstraction
  Technologies
• Step 5Integration of cue processing and context
  abstraction
• Step 6 Build applications

19
Step 1 Identify the contexts that matter

• The devise to be made smarter is to be analyzed.
• Is the devise used in changing environment?
• Do the expectations of the user toward the devise
  vary with situation?
• Is the interaction patter different in various
  situations?

20
Step 2 Find the appropriate Sensors

• For the variables found in Step 1, find
  appropriate sensor based on the following cues
• Accuracy of the sensor in relation to the
  variable
• Cost to provide the information

21
Step 3 Build and assess a prototypical devise

• Based on the sensors selected build a prototype
• Connect it to a standard data storage devise
• Test the data being accumulated
• Change sensors or data if needed

22
Step 4 Determine recognition and abstraction
technologies

• Identify a set of cues, with minimal data loss
• Select an algorithm that recognizes the context
  with maximal certainty and minimal data loss

23
Step 5 Integration of Cue processing and the
Context Abstraction

• Integrate the sensing technology and processing
  methods in a prototype
• May need to change algorithms, cues or Sensor
  based on the performance of the prototype

24
Step 6 Build the Application

• Build applications that use context knowledge

25
Cost Function

• Power consumption
• Size and weight
• Price of components
• Robustness and Reliability

26
Smart applications

• Applications behaving in accordance with their
  surroundings
• Examples
• Mobile user applications

27
Example Mobile Phone

• Mobile phones can switch from one configuration
  to another
• Example profile-to-profile
• Hand vibrating mode
• Table gentle sound
• Pocket Silent
• Outside Loud ringing
• General if no other context can be detected

28
Conclusion

• Smart Device is one that is not ignorant of its
  environment
• Information given by the environment can be used
  to build smart devices
• The context awareness can be used in the
  appliances like mobile phones.

29
Finally

• ANY QUESTIONS ?