狠狠撸

Data transformation and query management in personal health sensor networks

Mark Roantree a, Jie Shi a, Paolo Cappellari a, Martin F. O’Connor a, Michael
Whelan b, Niall Moyna b
a Interoperable Systems Group, School of Computing, Dublin City University,
Ireland.
b School of Health and Human Performance, Dublin City University, Ireland
Journal of Network and Computer Applications 35 (2012) 1191–1202

Article history: Received 30-July-2010, Received in revised form 27- February-2011, Accepted 10-May-2011, Available online 27-May-2011

Keywords: Wireless sensor network, Data synchronization, Calibration, Query service

Reporter
NILAMADHAB MISHRA
D0121008

& 2011 Elsevier Ltd. All rights reserved.

Contents

? Introduction
? Managing data in sensor network
? A method for outlier removal
? Data transformation and storage model
? Meta data services
? Experiments and results
? Conclusion

2

Introduction-1
? Here the authors address the issue of sensor network
technology supporting health care applications, where
sensors monitor human beings, the environment and provide
continuous streams of information.
? In particular, authors focus the attention on the case of sensor
networks provided by football teams with sensors generating
continuous heart rate values during a number of different
sporting activities.
? The requirement is for effective and efficient methods of
data management and transformation, in order to present
data in a format suited to high-level queries.

Cont.… 3

Introduction-2
? Research problem – To minimize the gap between the
low-level sensor output and the high-level user requirements
drawn by a traditional database-style query interface where
domain experts can continue to probe for the answers
required in more specialized environments.
? Proposed solution- The solution adopted by authors
consists in a novel approach that automatically harvests the
raw sensor data and provides semantic enrichment through
the addition of context data. Experimental assessment on
several study cases are provided and discussed in detail.

Cont. 4

Introduction-3
? The main motivational background to do this research-
Gaelic football is the most popular sport in Ireland. It is a hybrid
of Rugby and Australian Rules football. This project assessed heart rate
responses during small sided and regular Gaelic football games in
young players. To achieve this, author created a wireless sensor
network that has multiple configurations and requires a sophisticated
data management layer to process, normalize and query the data
streams.
? The aim is to determine if the maximal heart rate is attained by all
players while varying two parameters: the size of the playing area
and the number of players per team. The challenge is how to
capture and make sense of the large volumes of data that will be
gathered during each experiment.

5

Introduction-4

As heart rate data indicates the amount of time spent in different intensity zone , so
we use sensors to detect the heart rate during game time only. If a player’s heart
rate value is measured in every 5 second (during playing) then there are
approximately 10800 heart rate values generated by a team of 15 players during a
60 minutes game of Gaelic football .
So it’s a challenge to devolve a method to automatically harvest and transform the
low level sensor data , so as to meet the high level requirements. Here the authors
also try to provide a traditional query interface for low level data generated by WSN
and to process the data streams to enable an easier query expression. 6

Contents

? Introduction
? Conclusion

7

Managing data in sensor network
? Aim is to develop a data management layer in order to
process and calibrate (standardize/normalize) data generated
by WSN.

"Outliers" are values that "lie outside" the other values. When we collect data, sometimes there are values that are "far
away" from the main group of data.
Data enrichment is the concept of adding additional value to existing data, based on the some intents.(data quality
assurance)
Envelopes /covers sensor’s harvesting row data into XML and then stored in MONET DB XML data base and transform
the data into some standard format to meet high level requirements. 8

Managing data in sensor network
? Network topology is used for experiment. Sensed data are
transmitted to base station through heart rate monitors from
every participating node in the network.
? Each player belts the wireless device shortly in chest before
the activity begins and removes this quickly after the activity
ends.
? The purpose of structural enrichment process is to convert
the row sensor data into XML format with providing both
structural and meaningful semantics.
? The purpose of filtering process is to remove all the heart rate
values (outlier),those are not the part of spotting activity
because we want to analyze only the heart rate values during
spotting.

9

Contents

? Introduction
? Conclusion

10

A method for outlier removal
? Here the authors try to develop a new process for identifying
and eliminating the outlier values generated by sensors.
? In all sensor network the sensor will generate outlier values
which are clearly outside the normal acceptable range, so
before any query or transformation of data can progress, it is
required to detect and calibrate these outliers.
? Here authors propose a generic model that operates on XML
sensor output and can be parameterized by domain specialist.
? There are 4 primary steps -----
1*set valid range.
2*identify candidate outlier.
3*identify actual outliers.
4*calibrate outliers.
11

A method for outlier removal
1*set valid range.
(set valid range, min value, max value)In heart rate monitoring the general rule of
thumb for maximum heart rate limit =220bpm and minimum heart rate=78 bpm.
2*identify candidate outlier.
Here we examine all heart rates which are within the above defined valid range. So
any heart rate value within the above range is a candidate outlier.
3*identify actual outliers.
Read the 5HRs before and 5HRs after the candidate outlier. Calculate the mean of
these10 values (call this Mean Compare).
If the candidate outlier is outside 1.5% of Mean_Compare, (where 1.5% is a variance
parameter)
Then it is considered an actual outlier.
4*calibrate outliers.
Once consider an actual parameter, it is replaced with the mean-compare which is
calculated in step-3.

12

Contents

? Introduction
? Conclusion

13

Data transformation and storage model

Basic sensor schema S1 {endurance ..stamina} 14


Transformed schema S2 15


16

Contents
? Introduction
? Related research
? Conclusion

17

Meta data services

18

Meta data services
? Add new sensor- when domain users needs to introduce a new
sensor into an experiment , they simply provide a template file and
system can read the sensor data.
? Add new context-Here authors provides set of states- timing offset ,
timing interval etc.
? Apply activity-when a new sensor is introduced ,the template file
facilitates reading and processing the output.
? Add new transformation –it includes Transformation of the basic
schema to enhanced refined schema(S2) and mapping of
transformed objects.

19

Contents

? Introduction
? Conclusion

20

Experiments and results

MONET DB 21

Experiments and results

The standard query sets are provided by domain expert and these query have
been executed on sensor data.

The response time for each query includes the time required to execute the
query and returns the results to users.

MONET DB
22

Contents

? Introduction
? Conclusion

23

Conclusion
? Here the authors focused on how exercise scientist developed
a series of health sensor networks to monitor the physical
effects on different spotting activities on young children.

? The development of meta model and meta data service
facilitates the customization of transformation process.

? Suggests the automated process of transforming data
generated by sensors into a schema based format to provide
interface for a wide users community.

24

狠狠撸

Data transformation and query management in personal health sensor network

Recommended

More Related Content

Viewers also liked (11)

Similar to Data transformation and query management in personal health sensor network (20)

Data transformation and query management in personal health sensor network