Web-based Remote Human Pulse Monitoring System with Intelligent Data Analysis for Home Healthcare

Similar documents
Sinusoidal Steady State Analysis

Simplified Algorithm and Hardware Implementation for the (24, 12, 8) Extended Golay Soft Decoder Up to 4 Errors

VI.C CIRCUIT BREAKERS

Pre-distortion Linearization for 64-QAM Modulation in Ka-Band Satellite Link

Superposition, Thevenin and Norton. Superposition

Knowledge Unit Relation Recognition Based on Markov Logic Networks

A Comparison of South East Asian Face Emotion Classification Based on Optimized Ellipse Data Using Clustering Technique

A Novel Control Method for Input Output Harmonic Elimination of the PWM Boost Type Rectifier Under Unbalanced Operating Conditions

Math Circles Finite Automata Question Sheet 3 (Solutions)

Data Attribute Reduction using Binary Conversion

Efficient Large Integers Arithmetic by Adopting Squaring and Complement Recoding Techniques

Rough Set Approach for Categorical Data Clustering 1

An Optimal Method for Using Multiple Gateways in Cellular IP Networks

Semi-quantum private comparison protocol under an. almost-dishonest third party

COVERAGE HOLES RECOVERY ALGORITHM BASED ON NODES BALANCE DISTANCE OF UNDERWATER WIRELESS SENSOR NETWORK

2-7 Calibration of SAR Probe

Fixation-Image Charts

Energy Efficient Session Key Establishment in Wireless Sensor Networks

BnB-ADOPT + with Several Soft Arc Consistency Levels

EFFECTIVE CURRENT CONTROL DESIGN AND ANALYSIS OF SINGLE PHASE INVERTER FOR POWER QUALITY IMPROVEMENT

A photoplethysmograph experiment for microcontroller labs

MAXIMUM FLOWS IN FUZZY NETWORKS WITH FUNNEL-SHAPED NODES

Kirchhoff s Rules. Kirchhoff s Laws. Kirchhoff s Rules. Kirchhoff s Laws. Practice. Understanding SPH4UW. Kirchhoff s Voltage Rule (KVR):

Performance Evaluation of Survivable Multifiber WDM Networks

Control of assessment for demersal fish stocks in ICES area: analysis for 36 stocks and investigation of some potential bias sources.

i S1 V IN i C1 i N i C2 i S2

5 October 2015 Stereo Cross-feed Network for Headphones 1 of 12 Copyright 2015 Peter H. Lehmann. All Rights Reserved.

Rejection of PSK Interference in DS-SS/PSK System Using Adaptive Transversal Filter with Conditional Response Recalculation

A Substractive Clustering Based Fuzzy Hybrid Reference Control Design for Transient Response Improvement of PID Controller

Reinforcement Learning for Fuzzy Control with Linguistic States

Content Based Color Image Retrieval via Wavelet Transforms

Research on error compensation and measurement technology in robot flexible measurement

A Control Strategy Based on UTT and ISCT for 3P4W UPQC

IEE Electronics Letters, vol 34, no 17, August 1998, pp ESTIMATING STARTING POINT OF CONDUCTION OF CMOS GATES

Student Book SERIES. Patterns and Algebra. Name

Mixed CMOS PTL Adders

Multi-beam antennas in a broadband wireless access system

On The Study of Establishing a Responsive Infrastructure for a Massively Multiplayer On-Line Game

INSTITUTE OF AERONAUTICAL ENGINEERING (Autonomous) Dundigal, Hyderabad

Design of solar insolation power logging system with ADC error testing and correction

Design of Neuro-Fuzzy System Controller for DC Servomotor- Based Satellite Tracking System

Experiment 3: The research of Thevenin theorem

Optimal Placement of PMU and RTU by Hybrid Genetic Algorithm and Simulated Annealing for Multiarea Power System State Estimation

Discontinued AN6262N, AN6263N. (planed maintenance type, maintenance type, planed discontinued typed, discontinued type)

10.4 AREAS AND LENGTHS IN POLAR COORDINATES

A High-Sensitivity Oversampling Digital Signal Detection Technique for CMOS Image Sensors Using Non-destructive Intermediate High-Speed Readout Mode

FPGA Implementation of Ultrasonic S-Scan Coordinate Conversion Based on Radix-4 CORDIC Algorithm

Student Book SERIES. Fractions. Name

ABB STOTZ-KONTAKT. ABB i-bus EIB Current Module SM/S Intelligent Installation Systems. User Manual SM/S In = 16 A AC Un = 230 V AC

A Multi-Agent Solution to Distribution System Management by Considering Distributed Generators

Genetic Algorithm for Sensor Scheduling with Adjustable Sensing Range

FPGA Implementation vector control of tandem converter fed induction machine

The Math Learning Center PO Box 12929, Salem, Oregon Math Learning Center

UE Centric Coordinated Beamforming in Multi-cell MU-MIMO Systems

Chapter 7. Analysis of Variance

High Speed, Low Power And Area Efficient Carry-Select Adder

Modified Venturini Modulation Method for Matrix Converter Under Unbalanced Input Voltage Conditions

To provide data transmission in indoor

Solutions to exercise 1 in ETS052 Computer Communication

International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 INTERNATIONAL JOURNAL OF ADVANCED RESEARCH IN

An iterative approach to an integrated land use and transportation planning tool for small urban areas

1 tray of toffee 1 bar of toffee. 10 In the decimal number, 0 7, the 7 refers to 7 tenths or

LATEST CALIBRATION OF GLONASS P-CODE TIME RECEIVERS

Dynamic Optimization. Assignment 1. Sasanka Nagavalli January 29, 2013 Robotics Institute Carnegie Mellon University

Single-Phase SOGI-PLL Based Reference Current Extraction for Three-Phase Four-Wire DSTATCOM

DEVELOPMENT OF AN EFFICIENT EPILEPSY CLASSIFICATION SYSTEM FROM EEG SIGNALS FOR TELEMEDICINE APPLICATION

LOW SENSITIVITY DESIGN OF MULTIPLIERLESS ELLIPTIC IIR DIGITAL FILTERS

METHOD OF LOCATION USING SIGNALS OF UNKNOWN ORIGIN. Inventor: Brian L. Baskin

458 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 1, JANUARY 2016

The Discussion of this exercise covers the following points:

A New Type of Weighted DV-Hop Algorithm Based on Correction Factor in WSNs

Patterns and Algebra

VRT014 User s guide V0.8. Address: Saltoniškių g. 10c, Vilnius LT-08105, Phone: (370-5) , Fax: (370-5) ,

On the Description of Communications Between Software Components with UML

The Dynamic Utilization of Substation Measurements to Maintain Power System Observability

A Novel Back EMF Zero Crossing Detection of Brushless DC Motor Based on PWM

Figure 1. DC-DC Boost Converter

Exercise 1-1. The Sine Wave EXERCISE OBJECTIVE DISCUSSION OUTLINE. Relationship between a rotating phasor and a sine wave DISCUSSION

Software Pipelining for the Pegasus IR

CONTAINER BERTH SCHEDULING POLICY WITH VARIABLE COST FUNCTION

IMPROVEMENT OF THE QUALITY OF ELECTRICAL POWER BY A PHOTOVOLTAIC GENERATOR CONNECTED TO THE GRID

RC Filters TEP Related Topics Principle Equipment

Mitigation of Harmonics by Shunt Active Power Filter using Synchronous Detection Method

MEASURE THE CHARACTERISTIC CURVES RELEVANT TO AN NPN TRANSISTOR

Research on Peak-detection Algorithm for High-precision Demodulation System of Fiber Bragg Grating

熊本大学学術リポジトリ. Kumamoto University Repositor

Walsh Function Based Synthesis Method of PWM Pattern for Full-Bridge Inverter

IMPACT OF AIRPORT NOISE REGULATIONS ON NETWORK TOPOLOGY AND DIRECT OPERATING COSTS OF AIRLINES

How to Get Started with Virtual Goalie

Robot Deception: Recognizing when a Robot Should Deceive

Figure 1. DC-DC Boost Converter

Performance Analysis of Multi User MIMO System with Block-Diagonalization Precoding Scheme

A Development of Earthing-Resistance-Estimation Instrument

GLONASS Inter-frequency Biases and Their Effects on RTK and PPP Carrier-phase Ambiguity Resolution

PRO LIGNO Vol. 11 N pp

9.4. ; 65. A family of curves has polar equations. ; 66. The astronomer Giovanni Cassini ( ) studied the family of curves with polar equations

Safety Relay Unit. Main contacts Auxiliary contact Number of input channels Rated voltage Model Category. possible 24 VAC/VDC G9SA-501.

Priority based Dynamic Multiple Robot Path Planning

PRACTICAL, COMPUTATION EFFICIENT HIGH-ORDER NEURAL NETWORK FOR ROTATION AND SHIFT INVARIANT PATTERN RECOGNITION. Evgeny Artyomov and Orly Yadid-Pecht

A Simple Satellite Exclusion Algorithm for Advanced RAIM

Transcription:

We-sed Remote Humn Pulse Montorng System wth Intellgent Dt Anlyss for Home Helthcre Chh-Mng Chen Grdute Insttute of Lrry, Informton nd Archvl Studes, Ntonl Chengch Unversty, Tpe 6, Twn, R.O.C. chencm@nccu.edu.tw Chn-Mng Hong Grdute Insttute of Appled Electronc Technology, Ntonl Twn Norml Unversty, Tpe 6, Twn, R.O.C. t726@ntnu.edu.tw Ch-Meng Hung Grdute Insttute of Industrl Educton, Ntonl Twn Norml Unversty, Tpe 6, Twn, R.O.C. ntnu_dddrem@yhoo.com.tw T-Hung Lee Grdute Insttute of Lernng Technology, Ntonl Hulen Unversty of Educton, Hulen 97, Twn, R.O.C. mthp24@yhoo.com.tw Astrct Mny studes hd ndcted tht pplyng ntellgent systems wth physology sgnl montorng for e-helth cre s current developng trend. A physology sgnl montorng system cn help medcl stffs to montor nd nlyze humn s physology sgnl effectvely, such tht they cn not only montor the ptents physology sttes mmedtely, ut lso reduce medcl cost nd sve lot of tme of ptents to vst hosptl s doctors. Therefore, the study employed system on chp (SOC) technques to develop n emedded humn pulse montorng system wth ntellgent dt nlyss mechnsm for dsese detecton nd long-term helth cre, whch cn e ppled to montor nd nlyze humn s pulse sgnl n dly lfe. Menwhle, the proposed system lso developed frendly we-sed nterfce tht s convenent to the oservton of mmedte humn physologcl sgnls. Moreover, ths study lso proposes n ntellgent dt nlyss scheme sed on the modfed cosne smlrty mesure to dgnose norml humn pulses for explorng potentl chronc dseses. Therefore, the proposed system provdes enefts n terms of dng long-dstnce medcl tretment, explorng trends of potentl chronc dseses, nd urgent stuton nformng for sudden dseses. Keywords pulse physology sgnl montorng system, emedded system, e-helth cre, sequence dt nlyss I. INTRODUCTION Wth progress of medcl technologes nd development of newer nd etter medcnes, mny countres re grdully comng nto gertrc socetes due to rpd growth of the gng populton, such tht the requrement of home helth montorng hs een ncresngly rsed for securng ndependent lves of elder or ged ptents []. Prtculrly, dvncements of physologcl sensors, wreless communctons, nd nformton technologes hve resulted n the rpd development of vrous wellness or dsese montorng systems, whch enle extended ndependent lvng t home nd mprovement of qulty of lfe for ndvduls [2]. Mny studes [2] [3][4] hd ndcted tht montorng humn physologcl nformton n rel-lfe condtons s especlly useful n mngement of chronc dsorders or helthcre prolems. Importntly, performng long-term montorng process s helpful to provde follow-up dt to oserve long-term trends n the wellness or helth sttus. In the pst yers, mny reserches pd much ttenton on developng long-term helth-cre montorng systems [5][6], whch re requred to e compct, lghtweght, low power consumpton, low-cost nd comfortle to e used t nytme. Furthermore, they must e le to detect physologcl sgnls relly nd stly n vrous dsturnces. Most of the reserches for helth-cre montorng n out-of-hosptl envronment hve concentrted on helth montorng t home [2][3][5][6][7][8]. Furthermore, there re reserch trends towrds developng werle physologcl montorng systems tht cn mesure vrous o-sgnls nd provde helthcre servces to user usng e-helth technology [7][8]. Moreover, the concept of remote montorng vtl sgnl y we-sed nterfce [9][], whch s helpful to trnsmt mmedte physologcl prmeters to physcns who re cple of nterpretng the mesurements for ndvduls, s lso key ssue for physologcl sgnl montorng systems. It s not only convenent to ndvduls who re ctvely wllng to receve feedck regrdng hs/her wellness or dsese sttus nd to prtcpte n hs/her own cre, ut lso provdes the eneft of dstnce tretment. Among humn vtl sgnls, ssessment of humn pulse hs long een reserch re of nterest n the physology feld ecuse the stutons of humn pulse reflect mny helth sttes [][2]. Humn pulse s the rte t whch humn hert ets. Therefore, humn pulse s usully clled hert rte, whch s the numer of tmes your hert ets ech mnute. However, the rhythm nd strength of the hertet cn lso e noted, s well s whether the lood vessel feels hrd or soft. Chnges n your hert rte or rhythm, wek pulse, or hrd lood vessel my e cused y hert dsese or nother prolem. Bsed on the mportnce of montorng humn pulse, the study med to develop novel we-sed remote humn pulse montorng system wth ntellgent dt nlyss sed on physologcl sensor, SOC (system on chp), wreless communcton, World Wde We, nd ntellgent dt nlyss technologes, for wellness montorng nd homecre. The mn emphss ws on uldng hndheld SOC-sed humn pulse montorng system, whch ntegrted wth wred home Internet network to trnsmt physologcl sgnls to remote physologcl nformton dtse through wreless communcton nterfce on SOC pltform. The proposed system lso developed frendly we-sed nterfce tht s convenent to the oservton of mmedte humn 978--4244-674-5/8 /$25. 28 IEEE CIS 28

physologcl sgnls to support dstnce tretment. In ddton, ths study lso proposes n ntellgent dt nlyss scheme sed on the modfed cosne smlrty mesure to dentfy the smlrtes etween ny two humn pulse sequences. The proposed scheme s helpful to dgnose norml humn pulses for explorng potentl chronc dseses. II. SYSTEM DESIGN Ths secton ntroduces the system desgn sed on physologcl sensor, sgnl preprocessng, emedded system, wreless communcton, nd World Wde We technologes. A.. System Archtecture Ths study presents we-sed remote humn pulse montorng system whch s composed of three prts ncludng the pulse sgnl mesurng module, the ARM emedded system pltform nd the remote server. Fgure presents the system rchtecture of the proposed system. The pulse sgnl mesurng module ms to sense the humn pulse sgnls v the pezoelectrc sensor nd to perform the sgnl preprocessng process for flterng out nosy sgnl. To trnsfer humn pulse sgnls to the emedded system, the nlog humn pulse sgnls re frst trnsferred nto dgtl sgnls y nlog-dgtl converter, nd then trnsmt to the emedded system through serl trnsmsson devce under predefned ud rte. When the collected humn pulse dt meet the predetermned mount, the emedded system trnsmts the humn pulse dt to the remote server y wreless communcton. Fnlly, the remote server stores the sensng humn pulse sgnls nto humn physology dtse s well s shows the physology sgnls to medcl personnel through we nterfce. The we-sed nterfce provdes eneft n terms of on-lne remote medcl dgnoss for home-helth cre. Furthermore, ths study presents n ntellgent dt nlyss scheme sed on symolc-sed hmmng dstnce smlrty mesure to detect norml humn pulse for home-helth cre. B. The Employed Scheme for Computng the Humn Pulse Rte Snce the humn pulse sgnl s smlr to the electrocrdogrm (ECG) sgnl, the rel-tme QRS detecton lgorthm [3] ws ppled to compute the humn pulse rte n ths study. The lgorthm cn detect the poston of R-wve, nd then use the ntervl tme etween ny two R-wves to compute the humn pulse rte. Fgure 2 llustrtes the detled procedures of estmtng humn pulse rte. The computng procedures nclude dfferenttng, squrng, movng-wndow ntegrton, flterng out the sgnl elements tht ther voltge levels re 2/3 low thn the hghest voltge level, computng ntervl tme etween ny two R-wves, nd estmtng humn pulse rte. Next, ths study gves n exmple to expln how to estmte humn pulse rte y the employed rel-tme QRS detecton lgorthm. Frst, Fg. 3 shows the orgnl pulse sgnl otned from the sgnl preprocessng crcut. Bsed on the rel-tme QRS detecton lgorthm, the orgnl humn pulse sgnl must e frst dfferentted to otn the slope nformton. Fgure 4 depcts the dfferentted result of the humn pulse sgnl. After performng dfferentton operton, the humn pulse sgnl s squred to emphsze the slope of the R-wve. Fgure 5 shows the humn pulse sgnl fter performng squrng operton. The next process s to perform the movng-wndow ntegrton operton. The movng-wndow ntegrton ms to flter out the nose ner the R-wve. Fgure 6 revels the humn pulse sgnl fter performng movng-wndow ntegrton operton. To preserve useful sgnl elements for computng humn pulse rte, the other sgnl elements should e fltered out excludng R-wve. Fgure 7 shows the humn pulse sgnl tht flters out the sgnl elements tht ther voltge levels re 2/3 low thn the hghest voltge level. Fnlly, the humn pulse rte cn e computed ccordng to the tme ntervl etween ny two R-wves. The humn pulse rte cn e estmted y the followng formul: f h = 6 () td where f h s the humn pulse rte, nd t d s the ntervl tme etween ny two turnng ponts. Fg.. System rchtecture of the proposed we-sed remote humn pulse montorng system Orgnl humn pulse sgn Dfferenttng humn pulse sgnl Squrng humn pulse sgnl Performng movng-wndow ntegrton operton Computng hert rte Fg. 2. The computng procedure of humn pulse rte Flterng out the voltge level s 2/3 low thn the hghest voltge level Computng the tme ntervl etween ny two R-wves

2 4 6 8 2 55 5 45 4 35 3 25 2 5 Fg. 3. The orgnl humn pulse sgnl 9 8 7 6 5 4 3 2 2 4 6 8 2 Fg. 5. The humn pulse sgnl fter performng squre operton 45 4 35 3 25 2 5 5 9 8 7 6 5 4 3 2 2 4 6 8 2 Fg. 4. The humn pulse sgnl fter performng dfferentton operton 45 4 35 3 25 2 5 5 2 4 6 8 2 Fg. 6. The humn pulse sgnl fter performng movng-wndow ntegrton operton 2 4 6 8 2 Fg. 7. The humn pulse sgnl tht flters out the sgnl tht ts voltge level s 2/3 low thn the hghest voltge level C. The Proposed Modfed Cosne Smlrty Mesure for Identfyng Anorml Humn Pulse Sequence Frst, n order to dentfy norml humn pulse sequence usng the proposed modfed cosne smlrty mesure, the symolc trnsformton ws used to represent humn pulse sgnl n ths study. The mjor concept s to trnsfer the humn pulse sgnl wth numercl dt type nto the humn pulse sgnl wth symolc dt type y ssgnng correspondng symol to ech smplng humn pulse sgnl. The prmry dvntge of the symolc trnsformton s to reduce the computtonl complexty of dentfyng norml humn pulse sequence, ut ths trnsformton could lose the orgnl chrcter of humn pulse sgnl. Therefore, ths study employs the slope trnsformton to dentfy the orgnl fetures of humn pulse sgnl efore performng the symolc trnsformton. After trnsformng the humn pulse sgnl nto the symolc dt sequence, the proposed modfed cosne smlrty mesure s ppled to estmte the smlrty degree etween ny two symolc sequences for dentfyng norml humn pulse sequences. Next, the slope nd symolc trnsformtons re detled s follows: (A) Slope trnsformton In order to dentfy the fetures of humn pulse sgnl for trnsferrng n orgnl humn pulse sgnl nto symolc dt sequence, the slope trnsformton s frst operted efore performng symolc trnsformton n the study. Suppose tht the sequence S = { s, s2,..., s,..., s m } represents m smplng humn pulse sgnl ponts, nd the sequence T = { t, t2, t3,..., t,..., t m } s ech correspondng smplng tme. The slope trnsformton cn e performed y the followng mthemtcl formul: s s + S =, m (2) t+ t where S = { s, s 2, s 3,..., s,..., s m } stnds for the trnsformed humn pulse sequence, s s the th smplng humn pulse sgnl pont, t s the correspondng smplng tme of the th humn pulse sgnl pont, nd m represents the totl numer of smplng humn pulse sgnl ponts. (B) Symolc trnsformton After performng the slope trnsformton, the vred sttuses of humn pulse sgnl re successfully extrcted for symolc trnsformton. There re two fctors tht need to e consdered when performng symolc trnsformton: one s how mny symolc chrcters should e used, nd nother s how much nformton of the humn pulse sgnl could e lost under usng lmted symolc chrcters for symolc trnsformton. Actully, these two consdertons re trde-off ssue. More symolc chrcters hve to e used f more complete humn pulse nformton need to e preserved, ut such symolc trnsformton wll ncrese the computtonl complexty of dentfyng norml humn pulse sgnls. Conversely, usng less symolc chrcters wll lose some humn pulse nformton, ut t cn reduce the computtonl complexty of dentfyng norml humn pulse sgnls. To mke ll smplng ponts of humn pulse sgnl tht hve performed the slope trnsformton ecomng comprtve sequences to ech other, the z-score normlzton must e frst conducted efore performng the symolc trnsformton. The z-score normlzton s formulted s follows: s s s = ' σ (3) where s stnds for the th humn pulse sgnl wth z-score th normlzton, s s the humn pulse sgnl wth ngulr trnsformton, s s the men of the humn pulse sgnls wth ngulr trnsformton, nd σ ' s the stndrd devton of the humn pulse sgnls wth ngulr trnsformton. The humn pulse sequence tht performs the z-score normlzton hs three propertes: the men s zero, the stndrd devton s one, nd ll humn pulse dt form norml dstruton. TABLE I dsplys the cut-pont of the norml dstruton sed on vrous trnsformton symols. In

TABLE I, the cut-ponts prtton the norml dstruton functon s severl non-overlpped res wth equl prolty. For exmple, two cut-ponts.43 nd.43 prtton the norml dstruton functon nto three non-overlpped res, nd ech re hs /3 occurrence prolty. TABLE I THE CUT-POINT OF THE NORMAL DISTRIBUTION BASED ON VARIOUS TRANSFERRED SYMBOLS The numer of Cut-pont trnsformton symols 3 (-.43,.43) 4 (-.67,,.67) 5 (-.84, -.25,.25,.84) 6 (-.97, -.43,,.43,.97) 7 (-.7, -.57, -.8,.8, -.57, -.7) 8 (-.5, -.67, -.32, -.4,.4,.32,.67,.5) 9 (-.22, -.76, -.43, -.4,.4,.43,.76,.22) (-.28, -.84, -.52, -.25,,.25,.52,.84,.28) (C) The proposed modfed cosne smlrty mesure for dentfyng norml humn pulse sgnl To dentfy norml humn pulse sgnl, ths study employs the cosne smlrty mesure to compute the smlrty degree etween ny two humn pulse sequences. Suppose tht the unon symols set otned from the humn pulse sequence P nd the humn pulse sequence P contns totl of k symolc terms. The smlrty degree etween two humn pulse sequences mesured y the cosne mesure cn e formulted s k s s Sm( P, P ) = (4) 2 s = k k 2 s = = where P = ( s s s,..., s ) nd P = ( s s s,..., s ),,..., 2 k, 2,..., represent respectvely the symol vectors of the humn pulse sequence P nd the humn pulse sequence P, s nd s stnd for the symol weghts of the th symol n the humn pulse sequence P nd the humn pulse sequence P, respectvely. In Eq. (4), the correspondng symol weght n the symol vector s computed sed on the symol frequency tht ppers n humn pulse sequence. However, the cosne smlrty mesure cnnot dentfy the smlrty degree etween ny two humn pulse sequences well, ecuse t cnnot completely revel humn pulse symolc semntcs. For exmple, the smlrty degree of the humn pulse sequence ABCD nd DCBA mesured y cosne mesure s equl to snce oth the humn pulse sequences nclude the sme set of symolc terms. Ths smlrty degree evluted y cosne mesure mples tht oth humn pulse sequences re completely dentcl n terms of semntcs, ut they hve completely dfferent menngs from the perspectve of humn pulse sgnl. In other words, the humn pulse symol tht ppers t dfferent postons represents dfferent menng. k To consder symol s poston for promotng the ccurcy when dentfyng norml humn pulse sequences, the proposed hmmng dstnce ws ppled to modfy the cosne mesure for mesurng the smlrty degree etween two humn pulse sequences wth ncresed precson. The proposed hmmng dstnce for oth the humn pulse sequences cn e computed usng the followng formul: H m P P = where L ( s ) nd ( s ) ( s ) L( s ) = L (5), L represents respectvely the trnsformed locton ndex vlues of the th symol tht ppers smultneously n the symol sequences of oth the humn pulse sequence P nd humn pulse sequence P, m s the totl numer of symols tht pper smetneously n the symol sequences of oth the humn pulse sequence P nd the humn pulse sequence P. To consder oth smlrty degree nd symol sequence semntcs smultneously when dentfyng norml humn pulse sequences, the modfed cosne mesure s expressed s H P, P Msm( P, P ) = ( ) Sm( P, P ) Mx( H ) (6) where Mx (H ) represents the mxmum hmmng dstnce mong the humn pulse sequence P wth ll dentfyng humn pulse sequences. Next, ths study gves n exmple to expln how to compute the hmmng dstnce etween two humn pulse sequences. Suppose tht the humn pulse sequence P contns four symolc terms fter performng symolc trnsformton, nd the symolc term set of the humn pulse sequence P s represented s P = { s, s2, s4}. To compute hmmng dstnce, term set P must e trnsformed nto locton vector sed on the order n whch symols pper. For exmple, the trnsformed locton vector of symolc term set P s represented s L ( P ) = (,2,3,4 ) for the correspondng symolc term set P = { s, s2, s4}. Moreover, suppose tht the symolc term sets of nother two humn pulse sequences P nd P re c P = { s, s5} nd P c = { s, s5, s3}, respectvely. Smlrly, the trnsformed locton vectors for oth P nd P re c L( P ) = (,2,3) for symolc term sets P = { s, s5}, s well s L( P c ) = (,2,3 ) for symolc term set P c = s, s, }. Therefore, the hmmng dstnces { 5 s3 etween the humn pulse sequence events P nd P c cn e computed s P H, = + 3 2 = P P wth the news

H, = + 3 3 = P P c III. EXPERIMENTS Currently, the proposed we-sed remote humn pulse montorng system ws mplemented on n emedded system wth Lnux operton system. The Apche we server, PHP lnguge, nd MySQL dtse were employed to develop the server sde system. Ths secton revels the mplemented we-sed remote humn pulse montorng system nd demonstrtes the performnce of the proposed modfed cosne smlrty for explorng norml humn pulses. A. The Implemented We-sed Remote Humn Pulse Montorng Fgure 8 shows the mplemented we-sed remote humn pulse montorng system, whch ncludes the pulse trnsducer pezoelectrc sensor for sensng humn pulse sgnl, sgnl preprocessng crcut for flterng our nosy sgnl nd mplfyng wek humn pulse, ARM emedded system wth Lnux operton system for trnsmttng humn pulse nto we server y wreless networks, we server wth humn physology dtse for collectng mmedte humn pulse sgnl, osclloscope for verfyng the qulty of sensng humn pulse sgnl, nd power supply. Usng ths system, medcl stffs cn drectly oserve humn pulse sgnl y we rowser from ny plce. Ths provdes enefts n terms of long-dstnce medcl tretment, explorng trends of potentl chronc dseses, nd urgent stuton nformng for sudden dseses. To verfy the mplemented system cn correctly mesure humn pulse sgnls, Fg. 9 smultneously shows humn pulse sgnl on oth the we nterfce nd osclloscope. Except dsplyng mmedte humn pulse sgnl, the we nterfce cn lso dsply the correspondng hert rte. Ths study confrmed tht the humn pulse sgnl dsplyed on the osclloscope s the sme wth the humn pulse sgnl dsplyed on the we rowser. Fg. 8. The mplemented we-sed remote humn pulse montorng system Fg. 9. The humn pulse sgnl smultneously dsplyed on oth the we nterfce nd osclloscope B. Explorng Anorml Humn Pulse y the Proposed Modfed Cosne Smlrty Mesure To demonstrte whether the proposed modfed cosne smlrty mesure cn correctly explore norml humn pulses, ths study mesured fve ctul humn pulse sgnls from fve dfferent persons to compre ther smlrtes to ech other y the cosne smlrty mesure nd the proposed modfed cosne smlrty mesure. Ths experment ms to demonstrte tht the proposed modfed cosne smlrty mesure cn correctly explore two most smlr humn pulse ptterns to ech other from humn pulse physologcl dtse. Ths wll e helpful to judge whether sensng humn pulse s smlr to ny norml humn pulses stored n humn pulse physologcl dtse. Fg. shows fve humn pulse sgnls sensed from fve dfferent persons. To compute smlrtes etween ny two humn pulses, the symolc trnsformton must e frst performed. Ths study dopted ten symolc chrcters to represent the orgnl humn pulse s symolc dt sequence y the proposed symolc trnsformton process. TABLE II shows the result of symolc trnsformton. TABLE III nd IV show the smlrtes mesured y the cosne mesure nd the proposed modfed cosne mesure, respectvely. The expermentl result lsted n TABLE III ndctes tht the humn pulse (.e. dt ) s most smlr to the humn pulse 4 (.e. dt 4). However, Fg. ovously shows tht the humn pulse s most smlr to the humn pulse 3 (.e. dt 3) n ths cse, not the humn pulse 4. The expermentl result lsted n TABLE IV confrms tht the proposed modfed cosne smlrty mesure cn correctly dentfy ths prolem ecuse the poston semntcs of humn pulse sequence nd symolc ppernce frequency re smultneously consdered. IV. CONCLUSIONS Ths study presents We-sed remote humn pulse montorng system wth ntellgent dt nlyss sed on physologcl sensor, SOC (system on chp), wreless communcton, World Wde We, nd ntellgent dt nlyss technologes for home helthcre n dly lves. The proposed system not only cn correctly sense nd trnsmt humn pulse

sgnls to humn physologcl dtse n server sde, ut lso provde frendly we-sed nterfce tht s convenent to the oservton of mmedte humn physologcl sgnls nd hert rte to support dstnce tretment n clent sde. In ddton, ths study lso presents modfed cosne smlrty mesure, whch cn dentfy norml humn pulse sequences, to support explorng potentl chronc dseses. More mportntly, the proposed system ws successfully mplemented y emedded system technology. Ths provdes enefts n terms of developng compct, lghtweght, low-cost nd hnd-held physology montorng system. 7 6 5 4 3 2 dt dt2 dt3 dt4 dt5 9.5.5 2 2.5 3 3.5 4 4.5 Fg.. T The humn pulse sgnls sensed from fve persons TABLE II. THE SYMBOLIC TRANSFORMATION RESULTS OF THE HUMAN PULSE DISPLAYED IN FIG. Elements 2 3 4 5 6 7 8 9 2 3 4 5 6 7 8 9 2 2 22 Dt A G J J J J H E F G E E C B C B F G G G F E Dt2 B J I J J J I E H B E I B E B C E H E E E E Dt3 A G J J J I G E D E F C C B C C D D F F E E Dt4 E J J J J I H F H H E C B B C D E F E F E E Dt5 B I J J J H G E E H G D B D D D E E G D E G Elements 23 24 25 26 27 28 29 3 3 32 33 34 35 36 37 38 39 4 4 42 43 44 Dt D D D D E E E E F E E E E E F E D C A A B I Dt2 E E H C C C H C E E C E H I C C C C A A A E Dt3 F D D E D D E E E E F E E E E F D C A A B I Dt4 E E F F E D F E F G E F F F F F F E D A A A Dt5 E B D E E E D D G G D E E E E E E D A A B J TABLE III. THE SIMILARITIES MEASURED BY THE COSINE MEASURE Dt 2 3 4 5 Dt(R).44794.77683.79644.6292 2.44794.39682.4336.4349 3.77683.39682.699.47775 4.79644.4336.699.5388 5.6292.4349.47775.5388 TABLE IV. THE SIMILARITIES MEASURED BY THE MODIFIED COSINE MEASURE Dt 2 3 4 5 Dt(R).493.5494.446.35393 2.493.7637.2653.933 3.5494.7637.2796.996 4.446.2653.2796.44 5.35393.933.996.44 REFERENCES [] Stefnov D.H., Zeungnm Ben, Won-Chul Bng, The Smrt House for Older Persons nd Persons Wth Physcl Dsltes: Structure, Technology Arrngements, nd Perspectves, IEEE Trnsctons on Neurl Systems nd Rehltton Engneerng, vol. 2, ssue 2, pp. 228-25, 24. [2] Korhonen I., Prkk J., Vn Gls M., Helth Montorng n the Home of the Future, IEEE Engneerng n Medcne nd Bology Mgzne, vol. 22, ssue 3, pp. 66-73, 23. [3] Mlln J., Prk S.-E., Kefer S. nd Meyer J-U., TOPCARE-Implementton of Telemtc Homecre Pltform n Coopertve Helth Cre Provder Networks, The Second Jont EMBS/BMES Conference, pp. 869-87, 22. [4] Vrshney U., Mngng Wreless Helth Montorng for Ptents wth Dsltes, IT Professonl, vol. 8, Issue 6, pp. 2-6, 26. [5] Yu Y., Ou J., Desgn of Wreless Intellgent Sensor for Structurl Helth Montorng, Interntonl Conference on Intellgent Sensors, Sensor Networks nd Informton Processng, pp. -5, 4-7 Dec. 24. [6] Srovnl Vlem, nd Penhker Mrek, Helth Mntennce Emedded Systems n Home Cre Applctons, The Second Interntonl Conference on Systems (ICONS), pp. 7-7, 27. [7] Brnes G. & Wrren S., A Werle, Bluetooth-Enled System for Home Helth Cre, Proceedngs of the Second Jont EMBS/BMES Conference, vol. 3, pp. 879-88, 22. [8] Jnchu Yo, Schmtz R., nd Wrren, S., A Werle -of-cre System for Home Use Tht Incorportes Plug-nd-Ply nd Wreless Stndrds, IEEE Trnsctons on Informton Technology n Bomedcne, vol. 9, no. 3, pp.363-37, 25. [9] Kong K.Y., Ng, C.Y., Ong, K., We-Bsed Montor of Rel-Tme ECG Dt, Computers n Crdology, pp. 89-92, 2. [] Drs P., Bechtss D.K., Strntzs M.G., A We/WAP-sed System for Remote Montorng Ptents wth Dt Mnng Support, The 6th Semnr on Neurl Network Applctons n Electrcl Engneerng, 22. [] Lu O. nd Chwng A., Reltonshp etween Wrst-Pulse Chrcterstcs nd Body Condtons, In Proc. of the 4th Engneerng Mechncs Conf., 2. [2] Josh A.J., Kulkrn A. V., Chndrn S., Jyrmn V. K., Kulkrn B. D., Nd Trngn: A Pulse Bsed Dgnostc System, The 29th IEEE-EMBC Conf., pp. 227-22, 27. [3] Pn J. & Tompkns W.J., A Rel-Tme QRS Detecton Algorthm, IEEE Trnsctons on Bomedcl Engneerng, vol. 32, no. 3, pp. 23-236, 985.