Cis6930fa11_BBQ - Model Driven Data Acquisition in Sensor Networks Presented By Gautam S Thakur [email protected] Model Driven Data Acquisition

Info iconThis preview shows pages 1–15. Sign up to view the full content.

View Full Document Right Arrow Icon
Model Driven Data Acquisition in Sensor Networks Presented By: Gautam S. Thakur [email protected]
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Model Driven Data Acquisition in Sensor Networks • Authors – Amol Deshpande. – Carlos Guestrin – Samuel Madden • Presented at: – Very Large Database Conference, Toronto, in 2004
Background image of page 2
Is the Sensornet a „ Database’ ? • Well Yes and No, both – Yes • Contains data sources • Requires Declarative Querying – No • Not exhaustive • Sometimes no authoritative
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Two Main Problems [1] • Misrepresentation of Data – Impossible to gather all the relevant data – Data at discrete in space and time – Faulty sensors & misleading or erroneous values
Background image of page 4
Two Main Problems [2] • Inefficient Approximate Queries – Readings are approximate – Existing trends towards gathering as much data as possible – Query costs order of magnitude more than is appropriate
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Outline • Contribution • Approach • Proposed Model – BBQ • Model based Querying • Observation Plan – What and How to query ? • Experiments • Future Work and Extension • Conclusion & Q/A
Background image of page 6
Contribution • Using statistical models of real world processes • Why? – provide more robust interpretations of sensor readings (Account for biases, faulty or missing data) – A framework for optimizing sensor reading acquisition - when the model cannot provide an answer to the query.
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
• Key optimization problem: Acquisition plan for sensornet - given query and model, to get most refined answer • Two Dependencies: – Statistical benefits of getting a reading – vicinity based updates, improve the overall quality of results – System costs of getting a reading – connectivity of WSN
Background image of page 8
• Prototype: BBQ – Barbie-Q: A model based querying approach – Based on time-varying multivariate Gaussians – Generic model based architecture tailored for BBQ
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Overview of Approach • BBQ: – Declarative query processing engine using probabilistic model to answers queries about the current state of the sensor network • Model: – Probability density function (pdf) – p(X 1 ,X 2 ,…,X n ) – Probability assigned to each possible assignment to attributes X 1 ,…,X n where X i is attribute at particular sensor
Background image of page 10
• Users fire – SQL like queries – However, request real-time information about environment rather than stored data • Model is used to estimate sensor readings in the in real time (current) – May obtain values from sensornet, refine estimates, reflect changes in sensor values – Going into a full circle
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
BBQ Architecture
Background image of page 12
Basic Architecture • Qualities – error tolerances / confidence – Model finds most efficient way to answer query (sensornet or approximate processing) – Knowledge of sensor network allows model to create observation plan and send to network – Network collects readings, model updated, query answers generated
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Correlation in Value • Temperature and voltage are highly correlated • Model can use the changes in voltage to infer temperature changes
Background image of page 14
Image of page 15
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/09/2011 for the course CIS 6930 taught by Professor Staff during the Fall '08 term at University of Florida.

Page1 / 62

Cis6930fa11_BBQ - Model Driven Data Acquisition in Sensor Networks Presented By Gautam S Thakur [email protected] Model Driven Data Acquisition

This preview shows document pages 1 - 15. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online