CMSC828K: Sensor Data Management Data Streams

1
CMSC828K Sensor Data Management Data Streams

• Amol Deshpande

2
Today..

• Introductions
• Overview of the syllabus
• What ? Why ?
• Grading/Class requirements etc...
• No laptops in the class

3
Why ?

• Emergence of sensing devices that can be
  networked together on a large scale
• New data management challenges
• Very high-rate data streams
• Uncertain/imprecise data
• New types of queries
• We need to develop techniques to handle such data

4
A Sensor

• A device that can sense things
• sense measure instrument
• Examples
• Traffic Sensors
• e.g. traffic cameras
• Location Sensors
• e.g. cell phones, GPS units
• Sensors sensing environmental properties
• e.g. temperature, humidity, light etc

5
Sensor Network

• A collection of sensing devices that can
  communicate with each other
• Can collectively measure or instrument a large
  scale phenomenon or property
• Increasing number of deployments everywhere
• Fueled mainly by developments in MEMS

6
MICA2 Mote (Berkeley mote)
7
Types of Sensors

• Temperature
• Variable resistors that change resistance with
  temp
• Photocells, fog detectors
• MEMS Micro-electronic Mechanical Systems
• Acceloremeters, gyroscopes, tilt sensors
• Soil moisture (electrical resistance)
• GPS
• Acoustic (microphones)
• Cameras ?

8
Types of Sensors

• Wireless camera sensors ?
• Several projects underway
• CMOS based imaging sensors
• Very low power, but very low resolution

9
WSNs Vital Signs Monitoring
Project Codeblue Harvard
Can monitor the vital signs (for months on 2
batteries), and transmit it wirelessly. Can be
used to detect anomalous behavior, to raise
alarms etc
10
RFIDs (Smart Labels)

• Identify objects from distance
• small IC with RF-transponder
• Wireless energy supply
• 1m, magnetic field (induction)
• ROM or EEPROM (writeable)
• 100 bytes
• Cost 0.1 ... 1
• Consumable and disposable
• Flexible tags
• laminated with paper

11
RFIDs (Smart Labels)

• RFID Increasingly number of deployments
• Supply chain management
• Tracking Luggage, medical equipment, clothes ...
• Activity recognition
• RFID Tags iGlove

12
Wired Sensor Networks/Macro-scopes

• Traffic/GPS sensors, video cameras, microphones
• Disaster management, Surveillance, Tracking,
  Activity detection
• Somewhat different challenges
• Especially for audio/video sensors
• But a lot of commonalities...

Thanks to Gutemberg Bezerra
13
Applications..
Habitat monitoring Elder care Seismic structure
monitoring Home automation Contamination
tracking Measuring pollutants Location-based
services Traffic monitoring Supply-chain
management (RFID)
Precision Agriculture/Nursery Object
tracking Smart environments Surveillance Industria
l Monitoring Interactive Museums Battlefield
applications Swimming Pool Monitoring etc...
etc...
14
Challenges

• Hardware platforms
• Started with the vision of smart dust
• Power consumption still an issue
• Battery power doesnt obey Moores law
• Reliability
• Deployments in extreme conditions
• Need to autonomously deal with failures

15
Challenges

• Programming interfaces/abstractions
• Still too much variety in the platforms...
• Networking for wireless sensors
• Reliable multi-hop is tricky
• Inherently lossy channel
• Routing protocols, connectivity
• Must deal with mobility, failures
• Localization, synchronization

16
Challenges

• Security/Privacy
• A very important issue...
• Monitoring has been happening, and will continue,
  no matter what
• WSNs just make it significantly easier
• Who controls the data ? Who sees it ?

17
Challenges

• Finally...
• the focus of the class...

18
Data management challenges

• Data generated in real-time and continuously
  (distributed data streams)
• Traditional database systems have significantly
  more static data
• Tremendous amounts of data generated
• Much of it useless, but need to process all
• Must be processed immediately

19
Data management challenges

• Typically acquisitional environments
• Data is not gathered/sensed until asked for
• Should carefully decide what to acquire
• Required to conserve power as well as for sanity
• Changes query processing quite fundamentally
• Acquisitional QP

20
Data management challenges

• Raw data is inherently uncertain
• Lossy sensor, communication link failures
• Imprecise errors in the sensing
• Uncertain/probabilistic fundamental limitations
  in the way sensing is done
• Use of statistical models fundamental
• Probabilistic Databases ?

21
(No Transcript)
22
Data management challenges

• Need for real-time statistical modeling
• Event/anomaly/pattern detection
• Removing noise from the data
• Spatial/temporal biases in the data

23
Data management challenges

• Data provenance
• Being able to trace something back to its origins
• Data exploration and visualization
• Managing large-scale spatio-temporal datasets
• Data interoperability
• Data security and privacy

24
Data management challenges

• Combination of all these factors has made this a
  very challenging and exciting research area....
• Multi-disciplinary solutions required
• Databases Machine Learning Networking

25
Class

• Class based on reading papers...
• Sorta-classic papers exist for two topics
• Wireless sensor networks
• Data streams
• Not so much about probabilistic modeling,
  uncertain data etc
• Schedule on the web....
• Class Forum

26
Outline

• Overview of sensing technologies, hardware
  trends, applications 2-3 classes
• Declarative query processing in sensornets 2-3
  classes
• Data streams system design, query processing and
  optimization, adaptive query processing 7-8
  classes
• Probabilistic graphical models and their role in
  sensor data management 4-5 classes
• Uncertain, probabilistic databases 5-6 classes
• Tentative schedule

27
Class Structure

• Before each class, email reading summaries to me
• Include 828 Summary in the subject.
• Each class, presentation followed by discussion
  about the papers
• Most by me
• Some by you
• If you plan to attend regularly but are not
  enrolled, you should consider doing a
  presentation as well

28
Grading

• Summaries 10
• Participation Presentation 10
• Project 40
• Literature survey
• Intermediate progress report
• Final report presentation (or maybe poster)
• Homework/Exam 40
• Remember, this is a graduate class

29
Thats all...

• Questions ?

30
Wireless Sensing Devices
31
Examples of Wireless SNs