2/20/2013

1

Interactive Real-time Interface for Smart Health Monitoring and

Analysis

MT2012 2/20/2013 1

Master thesis of the University of Aizu

m5151161, Achraf Ben Ahmed Supervised by Prof. Abderazek Ben Abdallah

Adaptive Systems Laboratory, Master of Computer Science and Engineering

Graduate School of the University of Aizu, Japan

MT2012

• Background

• Motivation

• Research goal

• Interactive real time interface features

• Evaluation

• Conclusion and future work

Outline

2/20/2013 2

Background (1/2)

• Multicore SoCs became an opportunity to satisfy the

high requirement of data and computation intensive

applications such as bio-medical data processing.

• We previously proposed and developed a novel

embedded health monitoring platform based on

various efficient HW and SW techniques

(BANSMOM*)

• The system can be easily adapted to different

subjects or different signals of interest.

*A. Ben Abdallah, Y. Haga, K. Kuroda, An Efficient Algorithm and Embedded Multicore Implementation for ECG Analysis in Multi-lead Electrocardiogram Records, IEEE Proc. of the 39th he International Conference on Parallel Processing, San Diego, pp.99-103, Sept. 13-16, 2010.

MT2012 2/20/2013 3

2/20/2013 MT2012 4

Background (2/2)

Research Motivation • Existing interfaces drawbacks:

– Manually-intensive work flow for data acquisition, formatting, and visualization. No information related to the node (name, location….)

No recording time

ECG wave is illustrated only with number format

• No support for secure RT visualization

• Need for a more robust system for collection, visualization, and analysis of physiological data.

MT2012 2/20/2013 5

Research Goal

• Design and evaluation of a Interactive Real Time Interface for monitoring purpose :

• Real-Time support Algorithm for data visualization

- Efficient data recording algorithm

- Smart record management

- Multiple-Nodes support scheme

- User friendly

- Data adaptation to current BANSMOM hardware

MT2012 2/20/2013 6

2/20/2013

2

IRIT General Structure

MT2012 2/20/2013 7

Remote server

SUP 1

SUP 2

SUP n

Node 1

Node 2

Node n

BANSMOM Platform Interactive Real Time Interface

PHP

MySQL

IRTI Workflow

MT2012 2/20/2013 8

index.php

application

• The IRTI was developed following the MVC pattern: Model, View , Controller

• All requests are independent since they are managed separately by the creation of an application instance

Multiple Nodes can be managed in parallel without a negative impact on the overall performance.

Controller

Filters

action

View

layout

model

User request

1 2

3

4

5

6 7

8

9

IRI Development Phases

MT2012 2/20/2013 9

Peaks Visualization Receive Data

Node Selection

Data Retrieval

ECG Data Capturing and Storing Phase

Data Visualization

Phase

-Data Captured from different nodes -Data classified by recording time and lead number

-Supervisor interface access -Node selection -Display node information

-Visualization mode selection - Old request data - Live request data

MT2012

BANSMOM ECG Processing Distribution

2/20/2013 10

𝑇𝐸𝑇 :Total execution time (10 seconds sample data)

R𝑇 :Lead reading time

𝑃𝑃𝐷𝑇 :PPD processing time

𝑆𝑇 : Storing Time

𝑇𝐸𝑇 = 𝑅𝑇 + 𝑃𝑃𝐷𝑇 + 𝑆𝑇 Sample No 16265 , 10 seconds, 14 periods, 84 peaks

𝑇𝐸𝑇 = 3.608 + 2.144 + 0.001 = 5.753𝑠 𝑇𝐸𝑇𝑝 = 0.06𝑠

MT2012

RT Support with Buffer Insertion in DB

2/20/2013 11

• We assume for 1 peak - Insertion frequency 𝐼𝑓 - BANSMOM processing time 𝑇𝐸𝑇𝑝 = 0.06 second/peak - Database update DBup = 0.07 second/peak - Reading frequency 𝑅𝑓 (by Ajax call)

𝑅𝑓 = 𝐼𝑓 = 𝑇𝐸𝑇𝑝 + Dbup

• A buffering mechanism approach is used in the DB level to cover any additional delay

ECG Data Capturing and Storing Phase

MT2012 2/20/2013 12

if DB empty?

Yes

Capture data[day, i ,j]

Store data

No Check DB[day-3]

abnormality

Node i assignment to supervisor K

Go to data process phase

If data is normal

Yes

Delete DB[day-3]

No

day=day-3

2/20/2013

3

Node Selection and Data Retrieval Phase

MT2012 2/20/2013 13

Yes

Check If new data

No

Inform supervisor

Check pass correct

No

Yes

supervisor k login

Select node

secure connection (SSl connection)

Display interface

Display list node supervisor k

Display information and history

Visualize peaks process

Database Info every 1 min

Data Visualization Phase

MT2012 2/20/2013 14

Visualize peaks

Node i selection

Choose range time

Get peaks from DB

Request old data

Request live data

Process peaks by the chart

Ajax call to the corresponding table i

$.ajax ({ type:"get" , url: "<?php echo CController::createUrl('ecg/GetSensorsDataLive') ?>",

Get peak by order (i,j) from DB

Process peak by the chart for (var i = 0; i<dataFromPHP.length; i++) { if (!pointExists(newData, dataFromPHP[i][0])) { newData.push(dataFromPHP[i]); }

chart.series[0].addPoint( eval(point),true, false); setTimeout(requestData, 1000);

Display obtained peak

new data in node i

table

Yes

Increment index j

No

peak(i,j) != peak(i,j-1)

Yes

Inform user about no more new incoming

data

No

[{"name":“Lead_name","data":[[time1,value1],[tim2,value2]

………..

MT2012 2/20/2013 15

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time

1 Current Time

Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

Database Update Simulation

1 2012-12-01 06:00:00.281

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012 2/20/2013 16

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time

1 2012-12-01 06:00:00.001

Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

1 2012-12-01 06:00:00. 375

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012 2/20/2013 17

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time

1 2012-12-01 06:00:00.001

Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

3 2012-12-01 06:00:00.422 1496

1 2012-12-01 06:00:00. 422

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012 2/20/2013 18

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time

1 2012-12-01 06:00:00.001

Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

3 2012-12-01 06:00:00.422 1496

4 2012-12-01 06:00:00.578 39

1 2012-12-01 06:00:00. 578

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

2/20/2013

4

MT2012 2/20/2013 19

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

3 2012-12-01 06:00:00.422 1496

4 2012-12-01 06:00:00.578 39

5 2012-12-01 06:00:00.070 45

1 2012-12-01 06:00:00. 422 1 2012-12-01 06:00:00. 070

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012 2/20/2013 20

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

Last_retrieved_peak

Id Last_peak_time Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

3 2012-12-01 06:00:00.422 1496

4 2012-12-01 06:00:00.578 39

5 2012-12-01 06:00:00.070 45

6 2012-12-01 06:00:00.883 104

1 2012-12-01 06:00:00. 070 1 2012-12-01 06:00:00. 883

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012 2/20/2013 21

RT Support with Buffer Insertion in Live-Region

Information_lead

Id Peak_time Peak_value

.. ….. ….

Last_retrieved_peak

Id Last_peak_time

1 2012-12-01 06:00:00. 883

Remote server

PHP

MySQL 1 2012-12-01 06:00:00.281 95

2 2012-12-01 06:00:00.375 -825

3 2012-12-01 06:00:00.422 1496

4 2012-12-01 06:00:00.578 39

5 2012-12-01 06:00:00.070 45

6 2012-12-01 06:00:00.883 104

7 2012-12-01 06:00:00.984 -816

1 2012-12-01 06:00:00. 984

Database Update Simulation

1

1

2 2

3

Ajax Call

time_i > Tref

Yes

i=i=1

Tref=ti Get

peak

MT2012

Development & Evaluation Methodology

2/20/2013 22

• Development Environment - Php 5.3.13 - MySQL 5.6 - Apache 2.2.22 - YII MVC framework 1.1 - Netbeans IDE 7.1

• Evaluation Parameters - Developed tool - Security and vulnerability issues - Real time visualization - Code Complexity

MT2012

Developed IRTI

2/20/2013 23

Node List View

MT2012

Developed IRTI

2/20/2013 24

Old Records Visualization View

2/20/2013

5

MT2012

Security and Vulnerability Evaluation

2/20/2013 25

• Scan parameters: (Netsparker web vulnerability and security testing) - Remote Code Evaluation, SQL Injection, Cross-site Scripting,

HTTP Header Injection

Vulnerability & security Evaluation

Parameter Value

Scan duration 00:35:07

Total request 34313

Average speed 16.28 req/sec

Nbr of issues found

20

Critical issues 0

Informational issues

20

MT2012

RT Visualization Evaluation

2/20/2013 26

Live data While DB is Updating

MT2012

Code Complexity Evaluation

2/20/2013 27

Models Views controllers

Files

(.php)

Codes

(lines) 70 430 200

Code complexity summary

Ecg,Ecghistory,Ecginfo

Ecginfomodel,Loginform

Responsible,sensor,user

sensorinfo

Index,Contact,login

List,Info,live

EcgController

Although the proposed IRTI SW tool has required features with RT capability, the total code is small and only about 700 lines.

MT2012 2/20/2013 28

Conclusion

• Design and evaluation of a Interactive Real Time Interface for BANSMOM monitoring system.

• IRTI tool is scalable and can display RT data

• When evaluation its security and vulnerability, IRTI has zero critical issue.

• IRTI code is small and user friendly. The total code is only about 700 lines.

MT2012 2/20/2013 29

Future Work

• Currently my developed SW tool supports only ECG data RT monitoring.

• Further study and test should performed with other biomedical data.

• Efficient Compression and Encryption Algorithms should be investigated for more secure communication.

MT2012 2/20/2013 30

Thank you for your listening

2/20/2013

6

Backup Slides

2/20/2013 MT2012 31

2/20/2013 MT2012 33

// remove the following line when in production mode defined('YII_DEBUG') or define('YII_DEBUG',true); // include Yii bootstrap file require_once('path/to/yii/framework/yii.php'); // create application instance and run $configFile='path/to/config/file.php'; Yii::createWebApplication($configFile)->run();

2/20/2013 MT2012 34

Mysql configuration :my.ini key_buffer = 1G max_allowed_packet = 1M table_cache = 64 sort_buffer_size = 16M net_buffer_length = 8K read_buffer_size = 256K read_rnd_buffer_size = 512K innodb_buffer_pool_size = 16M innodb_additional_mem_pool_size = 2M innodb_log_file_size = 5M innodb_log_buffer_size = 8M innodb_flush_log_at_trx_commit = 1 innodb_lock_wait_timeout = 50

2/20/2013 MT2012 35

2/20/2013 MT2012 36

Placement of electrodes

2/20/2013 37

1 lead refer to the tracing of the voltage difference between two of the electrodes.

For example "lead I" is the voltage between the right arm electrode and the left arm electrode "Lead II" is the voltage between the right arm and the feet.

Ref: http://en.wikipedia.org/wiki/Electrocardiography

MT2012