research article a real-time location-based services

Research ArticleA Real-Time Location-Based Services System UsingWiFi Fingerprinting Algorithm for Safety Risk Assessment ofWorkers in Tunnels

Peng Lin1 Qingbin Li1 Qixiang Fan2 Xiangyou Gao1 and Senying Hu1

1 State Key Laboratory of Hydroscience and Engineering Tsinghua University Beijing 100084 China2 China Yangtze Three Gorges Group Corporation Beijing 100038 China

Correspondence should be addressed to Peng Lin celinpetsinghuaeducn

Received 27 December 2013 Accepted 17 February 2014 Published 17 April 2014

Academic Editor Ying Lei

Copyright copy 2014 Peng Lin et alThis is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

This paper investigates the feasibility of a real-time tunnel location-based services (LBS) system to provideworkersrsquo safety protectionand various services in concrete dam site In this study received signal strength- (RSS-) based location using fingerprintingalgorithm and artificial neural network (ANN) risk assessment is employed for position analysis This tunnel LBS system achievesan online real-time intelligent tracking identification feature and the on-site running system has many functions such as workeremergency call track history and location query Based on ANN with a strong nonlinear mapping and large-scale parallelprocessing capabilities proposed LBS system is effective to evaluate the riskmanagement onworker safetyThefield implementationshows that the proposed location algorithm is reliable and accurate (3 to 5 meters) enough for providing real-time positioningservice The proposed LBS system is demonstrated and firstly applied to the second largest hydropower project in the world totrack workers on tunnel site and assure their safety The results show that the system is simple and easily deployed

1 Introduction

Over the past decade there was a surge of accidents inunderground constructions worldwide [1ndash3] Consequentlysafety is a critical issue in construction industry especiallyfor the underground construction workplace Currently thehealth and safety management of the human on a largehydroelectric power construction site mainly depends on thecontractor supervision and owner inspection on an irregularbasis Without modern information technology and effectivemanagement this traditional method can hardly assureworkersrsquo safety The construction industry has the poorestHealth and Safety (HampS) records in anymajor industry [4 5]In China the probability of constructionworkers being killedand injured is higher than the average figures for all otherindustries with billion RMB Yuan of economic losses beingmeasured each year [1]

As a consequence some studies [6ndash13] have been focusedon the development of intelligent control systems adopt-ing advanced communication technologies and intelligent

algorithms as the means for developing new automatedmanage systems These are expected to be capable of provid-ing support toHampS tasks in various circumstances For exam-ple some advanced technologies such as intelligent responseand rescue systems have significantly contributed to thereduction of mining fatalities and accidents [14 15] A real-time tracking service forworkers equipment andmaterials isthe important information for various construction activitiessuch as safety management [15 16] material management[17] and work planning Accurate and reliable informationof workersrsquo location can lead to better decision making

Positioning system technologies can be divided into twocategories that is outdoor [18 19] and indoor [17 20ndash25] positioning The most popular and established outdoorpositioning system is Global Positioning System (GPS) [18]The current indoor positioning systems technology is WiFiradio frequency identification (RFID) [20] laser infraredand ultrasound Modern localization systems use varioustechniques and algorithm such as received signal strengthindicator (RSSI) [16] time of arrival (TOA) time difference

Hindawi Publishing CorporationMathematical Problems in EngineeringVolume 2014 Article ID 371456 10 pageshttpdxdoiorg1011552014371456

2 Mathematical Problems in Engineering

of arrival (TDOA) and angle of arrival (AOA) GPS is well-known to work independently (defined as a device that doesnot require any installation of technology on a constructionsite other than a device on the resource to position it)for tracking service [19] RFID technology [20] enables aseamless link between any physical tagged entity and thebusiness information infrastructure providing lightweightcomputational and communication capabilities Currentlythe RFID technology that is used for personnel localizationis actually only an attendance recording system rather thanthe real location tracking system Using it not the exact pointof the target but only an area with bigger scope than thetarget can be knownThis does not satisfy the requirement ofreal-time precision positioning When the accidents happenin the tunnels it will be very difficult to rescue the trappedworkers because of low positioning accuracyWireless sensornetworks (WSN) have attracted more and more researchinterest in tunnel applications for their advantages of self-organization low cost and high reliability [24 26] Wirelesssensor networks for location tracking would allow for a widedeployment of sensors across construction sites and as aconsequence a chance for ubiquitous computing capable ofimplementing even complex applications such as integratedproject monitoring to identify the real state of constructionsite execution

With the development of IEEE80211 technology WiFispreads all over the world Its coverage becomes wider andwider Although WiFi is not designed for positioning thesignal that access point (AP) or station regularly sends con-tains the information of RSS which provides the possibilityfor locating mobile station The academia and industry paygreat attentions to applyingWiFi technology to locate pointsRFID has a history of demonstrated ability and marketdominance yet it also has a key disadvantagemdashthe factthat it is nowadays populated with proprietary solutionsincluding expensive readers WiFi-based real-time locatingsystem (RTLS) has recently become just such an opportunitywith RTLS functions being handled by specialized softwareCompared with the existing positioning technology such asGPS cellular localization RFID and ZigBee the positioningbased onWiFi has the following advantages [11] (1) can workon different occasions such as indoor and outdoor providingthe possibility of the ubiquitous positioning (2) only dependson the existing WiFi network does not need to make anychanges and is of low cost which means the existing ITinfrastructure can be reused (3) effect of non-line-of-sight(NLOS) on WiFi signal is small even in the situations wherethere are obstacles With the advances in data mining withbig data wireless communication technologies and smartmobile devices a real-time LBS which provide personalizedservices based on usersrsquo location information and the grow-ing accumulation of industry knowledge have been widelyapplied in military transportation logistics constructionsite and so forth Realization of environment monitoringand worker localization in underground constructions playsan important role in construction and workerrsquos safety Inorder to safeguard the workers firstly need to know wherethey are then we can carry out effective implementation of

LBS that provides rich and extensible services especially afterthe catastrophic accidents and so the rescue team can reachthe disaster scene accurately and carry out the rescue worktimely by acquiring accurate location information Thus insome industry like underground mining installing localiza-tion system is enforced by Chinese Governmentrsquos law Inhydropower industry leading corporations likeThree GorgesCorporation not only actively adopt sensory technologies inmonitoring health of hydraulic structural engineering [7 911] which makes a great contribution to cracking control ofmass concrete [9] but also pay more and more attention toLBS system to protect workersrsquo safety and various services

In this study the positioning system based onWiFi usingRSS is introduced The LBS system provides data based onthe location of the mobile client and can be segmentedinto ldquopushrdquo and ldquopullrdquo models The risk management is alsoproposed This paper studies the feasibility of an LBS systemto provide workersrsquo safety and protection and various servicesin tunnels of high arch dam site The proposed system wastested and deployed inmain tunnels of Xiluodu arch dam site

2 Methodology and Model

21 Position Calculation Method Currently the methods ofmost positioning systems based on WiFi using RSS aredivided into two main categories trilateration algorithm [2728] and fingerprinting [25] technique Trilateration algorithmestimates the target position by measuring the distancebetween target and at least three known reference pointswhile fingerprinting technique gets the target location bymatching the fingerprint information which is the character-istic of signals

211 Trilateration Algorithm Trilateration [27 28] is amethod of calculating the coordinates based on geometrywhich can be trivially expressed as the problem of findingthe intersection of three spheres that involves a systemof quadratic equations [28] There are many algebraic andnumerical methods to solve this problem in 2D [29] andin 3D positioning [28 30] The precondition of a simplifiedgeometric algorithm is that estimated distances from a nodeto at least three of the anchor nodes are known This methodis utilized in the three anchor nodes as the intersection ofa circle centered at the position of the unknown node asshown in Figure 1 The coordinates of the three anchor nodes1198601 1198602 and 119860

3is known in advance as (119909

1198861 1199101198861) (1199091198862 1199101198862)

and (1199091198863 1199101198863) and the distances from these three anchor

nodes to node 119860 are 1198891198861 1198891198862 and 119889

1198863 If the coordinates of

node 119860 (119909 119910) are unknown then there will be the followingformulae [29]

radic(119909 minus 1199091198861)2

+ (119910 minus 1199101198861)2

= 1198891198861

radic(119909 minus 1199091198862)2

+ (119910 minus 1199101198862)2

= 1198891198862

radic(119909 minus 1199091198863)2

+ (119910 minus 1199101198863)2

= 1198891198863

(1)

Mathematical Problems in Engineering 3

A3

A1

A2

A

y

x

z

Figure 1 Schematic diagram of spherical trilateration

Equation (2) is derived from (1) and the coordinates of119860are calculated

[119909

119910] = [

2 (1199091198861minus 1199091198863) 2 (119910

1198861minus 1199101198863)

2 (1199091198862minus 1199091198863) 2 (119910

1198862minus 1199101198863)]

minus1

times [

[

1199092

1198861minus 1199092

1198863+ 1199102

1198861minus 1199102

1198863+ 1198892

1198861minus 1198892

1198863

1199092

1198862minus 1199092

1198863+ 1199102

1198862minus 1199102

1198863+ 1198892

1198862minus 1198892

1198863

]

]

(2)

WiFi positioning based on trilateration algorithm canbe divided into two phases distance and location Firstlythe target point receives RSS of three different specific APswhose positions are known and then it is converted intothe distances between the target and the corresponding APsin accordance with the transmission loss model of wirelesssignal Wireless signals are commonly affected by path lossshadow fading and so on in the transmission process Therelationship between receiving signal power and the distancecan be given by signal transmission loss model

The location of target point is calculated through thetrilateration algorithm namely the three APs are centersrespectively The distances between the target and the cor-responding APs are the drawn radiuses of three circles Theintersection of three circles is exactly the target point TheWiFi positioning based on the trilateration algorithm reliesheavily on known AP location information and accuratesignal transmission loss model However due to reasons suchas increasingly complicated electromagnetism environmentin tunnels of arch dam it is hard to rely on the signal trans-mission loss model Therefore the wireless location based ontrilateration algorithmhas difficulties in this implementationand it is used as an auxiliary means

212 Fingerprinting Technique RSS-based location finger-printing is based on statistical theory and proven industrypractice Location fingerprinting refers to techniques thatmatch the fingerprint of some characteristic of a signal thatis location dependent [31] Position fingerprint identificationdepends on the database of characteristics of the targetFigure 2 illustrates that its process is mainly divided intotwo phases offline phase or training phase and onlinephase or positioning phase The goal of training phase isto establish a location fingerprinting database Firstly thereasonable reference pointrsquos distribution needs to be selected

Locations Fingerprints

Location 1 Fingerprint 1


Fingerprint database

Fingerprint collecting

Off-line

Fingerprint matching

Estimated location

middot middot middot

Figure 2 Schematic flow of RSS fingerprint technique

It is ensured that they can provide enough information forestimating position accurately at the positioning stage Thenin each reference point we measure the RSS values fromdifferent APs in turn making the corresponding uniqueidentity (usually MAC address) and position informationof reference point recorded in the database Due to theenvironmental effect the strength of wireless signals is notstable In order to overcome the influence of unstable RSS onpositioning several measurements on each reference pointare usually collected and averaged [25]The accurate positionis increased with collecting data however it means morelabor consumed

In this study the received specific RSS is comprisedof existing data in the database The position is calculatedaccording to certainmatching algorithm119870-nearest neighboralgorithm is often used to compare data in fingerprintingsystem [16 25]119870-nearest neighbor (KNN) method is one ofthe simplest ways to determine the location of the unknownposition by using the fingerprint map This algorithm isa location fingerprinting method that considers 119870 CPs(calibration points) to calculate the approximate position ofthe target The idea is to compare the fingerprints in thefingerprint map to the observed measurements and to select119870 calibration points with the ldquonearestrdquo RSS values In theKNN approach [32] the vector is used as a measurement andcompared to the fingerprint map which includes only thesample averages Let the list [32]

1198712

119896= 1198751 1198752 119875

119896 (3)

be the list of calibration point coordinates (3) correspondingto the list of 119870 fingerprints

1198861119870

= 1198861 119886

119896 (4)

which satisfies

119889 (119910 minus 1198861) le 119889 (119910 119886

1) (5)

where 119886119894

isin 1198861119870

119886119895

isin 1198861119896

and the function 119889(sdot) is achosen distancemeasureThe Euclidean norm is widely used


The most common choice as a userrsquos location estimator 119909 isthe average of the coordinates of the119870 ldquonearestrdquo fingerprintsthat is

119909 =1

119896

119870

sum

119894=1

119875119894 119875119894isin 1198711119896 (6)

The estimator is a very restricted approach to compute thelocation estimation because the number of possible estimatesis always finite and is a function of the number of CPs Thelocation estimation is done by using the value 119870 = 1 whichleads to the nearest neighbor (NN) method The Euclideannorm is used as a distance measure but the estimate isrejected if

10038161003816100381610038161003816119910119895minus 119886119894119895

10038161003816100381610038161003816gt 2119894119895 (7)

where CPi is the ldquonearestrdquo calibration point

22 Risk Management on Worker Safety Many studies[1 2 33] relative to risk management are conducted forunderground mining industry In China the probability ofconstruction worker fatalities in underground industries ishigher than that of all other industries with the economiclosses being measured each year in billions [1] Similarsituations have been recorded worldwide Health and safetycontrol are inadequate in terms of preventing risks whichbecause of their specific nature are unpredictable To betterassure the HampS of people in construction such sole depen-dence on patrolling officers for control purposes should bereduced and in part replaced by amore objective evaluation ofworker effort Techniques such as the tracking of the locationof workers and analysis of workersrsquo behavior would be ahelpful site control inclusion [8]Worker safetymonitoring atdam site is a complex multi-index nonlinear process whichrequires monitoring system has some intelligent informationprocessing capabilities in order to ensure the reliability Therisk management is also very important to proposal classifi-cation of safety risk of worker equipment and environmentand so forth Artificial neural network (ANN) [34] has astrong nonlinear mapping large-scale parallel processingcapabilities as well as adaptive self-training self-learningself-organization and fault tolerance and so forth It issuitable to be adopted in this LBS system

As shown in Figure 3 the 3-layer model [35] is the mostwidely studied and applied model among many differenttypes of artificial neural networks The first layer has inputneurons which send data via synapses to the second layerof neurons and then via more synapses to the third layerof output neurons More complex systems will have morelayers of input neurons and output neurons The synapsesstore parameters called ldquoweightsrdquo manipulate the data in thecalculations In this study the output represents the solutionto the problem that is worker safety assessment or risk index

Where (119883119899) is the input and (119882

119899) is the corresponding

array of weights the activation layer is given by

119886 =

119899

sum

119894=1

119883119894119882119894 (8)

Output

Output layer

Hiddenlayer

Inputlayer

I1

I2

In

Figure 3 The architecture of neural network model

Finally the output value can be calculated as

119910 = 119891 (119886 minus 120579) (9)

In this model the weight revision method can greatlyaffect the network behavior which is shown in the follwoing

119882119895ℎ(119905 + 1) = 119882

119895ℎ(119905) + 120578

120597119864

120597119881119895ℎ

+ 119886 lfloor119882119895ℎ(119905) minus 119882

119895ℎ(119905 minus 1)rfloor

119882ℎ119905(119905 + 1) = 119882

ℎ119905(119905) + 120578

120597119864

120597119881ℎ119905

+ 119886 [119882ℎ119905(119905) minus 119882

ℎ119905(119905 minus 1)]

(10)

where 119886 is the impulse coefficient 119882119895ℎ

is the connectedweight between the input layerrsquos node and the middle layerrsquosnode and 119882

ℎ119894is the connected weight between the middle

layerrsquos node and the output layerrsquos node

3 Description of the Real-Time LBS System

31 Overview of System This system focuses mainly on pro-viding a real-time LBS system that can automatically capturetransfer and analyze the positions of workers inworking zoneand can provide various location based services Figure 4shows the system main structural (1) fixed reference nodesin tunnels which are normalWiFi APs but the mechanics arechanged to fit for tunnel environments (2) installed gateway(3) moving nodes are Android smart phones running mobileLBS client (4) router firewall (5) database server (6) clientof showing map and workersrsquo realtime (7) location serverof running RTLS engine (8) location app server of runningsafety management software (9) GIS server which servesdigital map (10) fingerprint data server which is a fingerprintdata store

The systemrsquos main functions include the following

(1) Workerrsquos positioning service workerrsquos position datacan be stored persistently utilizing any moderndatabases such as Oracle and Microsoft SQL ServerOther systems can use the positioning service viastandard web service Thanks to load balancing tech-nology this system is designed to be flexible and hasa quick response even in case that the demands are


Client (web mobile)

Location server

WiFi AP

Smart phone

Location app server

GIS server

Fingerprint server

Database server

Routerfirewall Gateway

Figure 4 Component overview of the real-time monitoring system

in a heavy load Particularly the system can real-timehandle large amounts of workerrsquos position on-site

(2) Bidirectional alarm and warning service worker canissue critical alarm actively by pressing emergencybutton of tag control center can show the region andthe warning information When a worker goes intothe zone where it is dangerous or is forbidden alarmmessage will be alerted to the worker as well as hisworker mates or the supervisor near him

(3) Risk management safety issues are collected andmanaged by risk management module in LBS systemAn important part of issue tracking is to classify issuesas per their status LBS system assumes that an issuecan be in one of the three stages opened resolvedand closed Hence all listed statuses will be matchedto these three stages The following statuses arepredefined new acknowledged confirmed assignedresolved and closed

LBS system provides data based on the location of themobile client and can be segmented into ldquopushrdquo and ldquopullrdquomodels The ldquopushrdquo model is the one in which informationis proactively sent to subscribers On arch dam constructionsite there ismuch ofwarning and notice information needingto be delivered to works in certain areas in time so ldquopushrdquomodel is very useful ldquoPullrdquo services are used by subscribersto retrieve area information Workers can get informationfrom LBS according to their interest In order to strengthenthe robustness of data transmission the application offers atemporary storage function if communication with the basestation is broken The data is able to be uploaded when thenetwork is again available

Performance of the location-based service system in siteapplication is very important especially where there arehundreds of workers The following technical means are uti-lized for contributing high performance in this decentralizedand scale-out system (1) 3 up-to-date IBM xSeries serverspowered by multiway multicore Intel Xeon processors withhyper thread enabled (2) main data (system users workerrsquosinformation and WiFi finger print) load into memory andmain task happed there instead of store in Disk and avoidof frequently IO exchange A high-efficient key-value searchsubsystem which is similar to nowadays NOSQL database isdeveloped

32 Software Architecture This is typical client-server archi-tecture The service side is mainly responsible for locationrequest from terminal and positioning calculation Consider-ing the load balancing the web server and positioning serverrunning the position calculation are separate both logicallyand physically The client is mainly responsible for gatheringaround AP wireless signal strength and will submit thosedata to the server side the server using the data to calculateterminal position based on predefined algorithm

Using the standard HTTP client and server communi-cation protocol programming is convenient and scalableFigure 5 illustrates the information interaction diagram ofproposed software system Mobile terminal submits GETrequest to web server GET request information includessignal characteristic strength vector Web server receivesthose requests and forwards them to positioning serverPositioning server queries fingerprint database and doescertain calculation A best estimation for position worksout


Terminal client Web server Positioning server Database server

Get

GetGet

Result

Result

Result

Figure 5 Information interaction diagram of proposed software system

Monitoring zone

Organizationinformation

User management Worker management

Nine modules

Tunnel map and workertracking view

Worker information

Statistics andreport analysis

Figure 6 LBS web client interface

33 The LBS Software Interface The LBS software includesserver software web client and mobile interface Detailedintroduction is illustrated as follows

Server Software The server software consisted of (1) loca-tion engine server to calculate real-time position of mobileterminals using fingerprint mapping algorithm (2) mobileterminals management for management of all the mobileterminalsrsquo configuration diagnosing functions (3) systemadministration for user management system level parame-ters and so forth (4) data import and export for backupand restore (5) web mapping service (WMS) standardscompliant map server store sand displays spatial data Anyclient can use map service by embedding JavaScript snippetinto standard html page (6) log and diagnosis the entire logand diagnose information can be configured into differentcatalogs and levels the output destination can be selectablefrom local disk file TCPIP socket to restful web services(7) fault tolerance and load balancing one serverrsquos faultcannot lead to failure of whole system fault server can bedetected and isolated from the whole system system loadcan be distributed into servers according to resource usage

(CPU memory disk etc) (8) enterprise message serversystem and user-defined messages including warning andalert information to workers are delivered and dispatchedBoth SMS and LBS system message are supported

LBS Web Client Software Interface As shown in Figure 6 theproposed LBS web client interface consisted of nine modules(1) usermanagement whichmanages system users includinga RBAC based rights management (2) worker managementwhich manages all workers under LBS systemrsquos monitor-ing (3) virtual electrical fence system which can monitorworkers in and out certain area (4) map view web GISsystem which shows digital map (5) alarm system whichcollects all system alarms and notifications (6) attendancemanagement system which generates reports and analyzesthe working time sheet of all workers (7) report systemwhich reports and queries module for LBS (8) SMS systemwhere SMS can be sent out automatically or manually(9) safety issuesmanagement whichmanages all safety issuessuch as unsafe worker behavior and unsafe facility in workarea


(a) One tunnel view (b) WiFi AP (c) Wireless base station

Figure 7 A snapshot of RTLS devices in tunnel site

Table 1 The three-main-factor matrix

Risk probability (L) Risk frequency (E) Factor weight (C)

ScoreProbability of anaccident or dangerousoccurrence


Score Probability consequence

10 Absolutely possible 10 Always 100 Tragedy Many death6 Possible 6 Everyday 40 Very severe Some death

3 Possible but not often 3 Every week oroccasionally 15 Severe Loss of life or injured

seriously1 Almost not possible 2 Every month 7 Serious Injured

05 Possible rarely 1 Every year 3 Normal Injured slightly andcan be recovered

02 Hardly possible 05 Hardly happen 1 Concern Not healthy basically01 Impossible in realty

LBS Web Client software interface is in modern UI stylePopular web browsers including IE 80+Mozilla Firefox andGoogle Chrome are tested and verified

Mobile Client Interface The typical mobile GUI applicationalso is consisting of nine modules (1) location search whichprovides search function (2) web client which can invokemobile web browser (3) warning and notice which mainlyshows all the warnings and notices from system (4) systeminformation which shows current system status such as CPUload and memory usage (5) setting where reporting intervalvalue can be set server IP address can also be set (6) controlcenter which can use some utility of control center for exam-ple to broadcast an SMS to all users (7) 3G communication(both voice and video call) (8) emergency help which whentouched control center will receive an emergency message(9) help where help and tutorial information are provided

4 On-Site Application Cases

41 Site Setup Xiluodu hydroelectric power station [36 37]is the second largest one which can output 1386 million kWpower and it is close to Three Gorges hydroelectric powerstation in ChinaThe project site is located on the JinshajiangRiver in Leibo county of Sichuan ProvinceThe total pouringconcrete is about 600 million cubic meters and total lengthof tunnel is about 100000m There are hundreds of workersworking in the tunnels of arch dam in order to control

construction quality of dam grouting dam reinforcementand workerrsquo safety [8] The implementation of the proposedsystem is of important significance At 2013 WIFI-RTLSinfrastructure was deployed in all six main tunnels in archdam The installed snapshot is shown in Figure 7 In thetunnels site there is already complete WiFi network whichis used as backbone network for monitoring equipment anddevices to be used as communication tools

42 Discussion on Workerrsquos Trajectory and LBS Service RiskThe system monitored the workerrsquos activities in the tunnelTrack lines illustrated that the worker was doing his routineon-site safety supervision job while in the middle of tunnelthere is a place which is classified as ldquodangerousrdquo So LBSsystem pushes an alert message to him which is shown inFigure 8(c) This case shows the feasibility of a real-timemonitoring system for workerrsquos trajectory which functionsby checking whether any worker operates within a classifiedplace (work forbidden or dangerous area) If emergencysituations happened the worker can trigger an emergencycall for help On the other hand the system can pushmessageto worker By using this bidirectional communication LBSsystem functions are demonstrated andworkerrsquos on- siteworksafety is ensured

43 Discussion on Safety Risk Management Table 1 demon-strates current main risk factor matrix using Xiludu arch


Drainage tunnel

Grouting tunnelDrainage tunnel

Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform

(a) Behavior trajectory

669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079

RECORD ID COORD X COORD Y SYSTEM TIME5656426

5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723

middot middot middotmiddot middot middot middot middot middotmiddot middot middot

5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906

(b) Detail record of trajectory information

Close to the elevator shaft please pay attention to safety

(c) Alert message to worker

Figure 8 Workerrsquos trajectory process

dam construction site The table includes three-main-factormatrix risk probability (119871) risk frequency (119864) and factorweight (119862) The different score represents various probabil-ities of an accident or dangerous occurrence

The static risk assessment basis of following equation

Risk (119863) = 119871 times 119864 times 119862 (11)

There are twenty categories and many items in catalogfor source of risk in dam site Table 2 shows risk assessmentand regulation in tunnel site Every day total risk assess-ment is analyzed following the methodology as illustratedin Section 22 and an SMS (short message) is sent out tobe subscribed via the LBS system to deliver the summarymessage about risk management

The internal mechanism of ANN (artificial neural net-work) module of LBS system includes two stages one isldquotrainingrdquo and the other is ldquoapplyingrdquo As Figure 9 illustratedin the training stage historical data which was collected priorto the application of the LBS system was input into ANNmodules They are in 20 categories and more than 200 itemsaccording to Section 22 119883

1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882

119899from history data (paper record or Excel files)

Table 2 Risk assessment and regulation

Score Actiongt320 Dangerous all on-site operations need to be stopped160ndash320 Very dangerous need change immediately70ndash160 Dangerous need change20ndash70 Possible dangerous pay attentionlt20 A bit dangerous it is acceptable

are extracted and normalized The corresponding output(safety assessment) is also normalized So after training stageall the parameters such as 119886 and 120579 are generated from 119873

(total days of history data) times of iteration computing Themore the data the more accurate 119886 and 120579 Once the ANN isdone and ready to work some of the history data is kept forverification purpose then any new input which are gatheredfrom LBS data acquiring subsystem can get the output datausing the trained ANN Thanks to the LBS system there isa trend showing total safety situation for workers which aregetting better and better The conclusion is supported by thefollowing risk statistics diagram see Figure 10


Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value

Figure 9 Training flowchart for ANN module

05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29

Figure 10 Risk assessment of Xiluodu tunnel in October 2013

5 Conclusions

Thepaper presents the realization of real-time LBS system formonitoring workerrsquos location with the use of WiFi trackingtechnology to provide service base on the location Basedon the study results the most influential factors contributingto the successful implementation of the real-time LBS forworkers are identified

To achieve an online real-time intelligent tracking iden-tification feature the on-site running system satisfies workeremergency call track history and location query and soforth Based on ANN with a strong nonlinear mappingand large-scale parallel processing capabilities proposed LBSsystem is effective to evaluate the risk management onworkerrsquos safety

The site operation case also shows that the RSS-basedlocalization algorithm implemented by WiFi RTLS is reli-able and accurate enough in some cases but in other fewcases which require more accurate (less than 1m at cmlevel) positioning WiFi RTLS is not the final solutionSo hybrid positioning technology which includes differentprecision measurement needs to be developed on arch damconstruction site and more further researches need to beconductedMoreover LBS is in rapid development nowadaysboth in industry and in academia especially in 3D virtual

reality environment It can provide more vivid and perfectexperience to arch dam construction management firm

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This research work was supported by National NaturalScience Foundation of China (nos 11272178 and 51339003)National Basic Research Program of China (973 Pro-gram) Grant nos 2013CB035902 and 2011CB013503 andTsinghua University Initiative Scientific Research ProgramThe authors are very grateful to the ChinaThree Gorges Cor-poration for allowing access to one of its construction sites

References

[1] Q-H Qian and X-L Rong ldquoState issues and relevant recom-mendations for security risk management of Chinarsquos under-ground engineeringrdquo Chinese Journal of Rock Mechanics andEngineering vol 27 no 4 pp 649ndash655 2008

[2] A Asfaw C Mark and R Pana-Cryan ldquoProfitability andoccupational injuries inUS underground coalminesrdquoAccidentAnalysis amp Prevention vol 50 pp 778ndash786 2013

[3] P Berest ldquoAccidents in underground oil and gas storages casehistories and preventionrdquo Tunnelling and Underground SpaceTechnology vol 5 no 4 pp 327ndash335 1990

[4] S X Zeng VW Y Tam andCM Tam ldquoTowards occupationalhealth and safety systems in the construction industry ofChinardquo Safety Science vol 46 no 8 pp 1155ndash1168 2008

[5] J P Reyes J T San-Jose J Cuadrado and R SancibrianldquoHealth amp Safety criteria for determining the sustainable valueof construction projectsrdquo Safety Science vol 62 pp 221ndash2322014

[6] C Alessandro G Alberto and N Berardo ldquoA proactive systemfor real-time safetymanagement in construction sitesrdquoAutoma-tion in Construction vol 20 no 6 pp 686ndash698 2011

[7] P Lin Q-B Li and H Hu ldquoA flexible network structure fortemperature monitoring of a super high arch damrdquo Interna-tional Journal of Distributed Sensor Networks vol 2012 ArticleID 917849 10 pages 2012

[8] P LinQ-B LiQ-X Fan andX-YGao ldquoReal-timemonitoringsystem forworkersrsquo behaviour analysis on a large-damconstruc-tion siterdquo International Journal of Distributed Sensor Networkvol 2013 Article ID 509423 10 pages 2013

[9] P Lin Q-B Li S W Zhou and Y Hu ldquoIntelligent coolingcontrol method and system for mass concreterdquo Journal ofHydraulic Engineering vol 44 no 8 pp 950ndash957 2013

[10] T-H Yi H-N Li and X-D Zhang ldquoSensor placement onCantonTower for healthmonitoring using asynchronous-climbmonkey algorithmrdquo Smart Materials and Structures vol 21 no12 Article ID 125023 12 pages 2012

[11] B Naticchia M Vaccarini and A Carbonari ldquoA monitoringsystem for real-time interference control on large constructionsitesrdquo Automation in Construction vol 29 pp 148ndash160 2013

[12] T-H Yi H-N Li and M Gu ldquoOptimal sensor placement forstructural health monitoring based on multiple optimization


strategiesrdquo The Structural Design of Tall and Special Buildingsvol 20 no 7 pp 881ndash900 2011

[13] Q-B Li and P Lin ldquoDemonstration on intelligent damrdquo Journalof Hydroelectric Engineering vol 33 no 1 2014

[14] K-F Zhang M Zhu Y-J Wang E-J Fu and W CartwrightldquoUndergroundmining intelligent response and rescue systemsrdquoProcedia Earth and Planetary Science vol 1 no 1 pp 1044ndash10532009

[15] S Q Wang and Y Shi ldquoDesign of personnel position systemunder the mine based on RFID andWiFirdquo in Proceedings of theInternational Conference on Remote Sensing (ICRS rsquo10) pp 253ndash256 Hangzhou China October 2010

[16] X-W Luo W J OrsquoBrien and C L Julien ldquoComparative eval-uation of Received Signal-Strength Index (RSSI) based indoorlocalization techniques for construction jobsitesrdquo AdvancedEngineering Informatics vol 25 no 2 pp 355ndash363 2011

[17] N R Saiedeh and M Osama ldquoGPS-less indoor constructionlocation sensingrdquo Automation in Construction vol 28 pp 128ndash136 2012

[18] A Ibrahim and D Ibrahim ldquoReal-time GPS based outdoorWiFi localization system with map displayrdquo Advances in Engi-neering Software vol 41 no 9 pp 1080ndash1086 2010

[19] A H Behzadan Z Aziz C J Anumba and V R KamatldquoUbiquitous location tracking for context-specific informationdelivery on construction sitesrdquoAutomation in Construction vol17 no 6 pp 737ndash748 2008

[20] P Najera J Lopez and R Roman ldquoReal-time location andinpatient care systems based on passive RFIDrdquo Journal ofNetwork and Computer Applications vol 34 no 3 pp 980ndash9892011

[21] A Mkhida J-M Thiriet and J-F Aubry ldquoIntegration of intel-ligent sensors in Safety Instrumented Systems (SIS)rdquo ProcessSafety and Environmental Protection vol 92 no 2 pp 142ndash1492013

[22] G Deak K Curran and J Condell ldquoA survey of active and pas-sive indoor localisation systemsrdquo Computer Communicationsvol 35 no 16 pp 1939ndash1954 2012

[23] D Fahed and R Liu ldquoWi-Fi-based localization in dynamicindoor environment using a dynamic neural networkrdquo Interna-tional Journal of Machine Learning and Computing vol 3 no 1pp 127ndash131 2013

[24] R Alessandro C Marco B Luca C Matteo and M Tagliasac-chi ldquoAn integrated system based on wireless sensor networksfor patient monitoring localization and trackingrdquo Ad HocNetworks vol 11 no 1 pp 39ndash53 2013

[25] R Setiya and A Gaur ldquoLocation fingerprinting of mobileterminals by using Wi-Fi devicerdquo International Journal ofAdvanced Research in Computer Engineering amp Technology vol1 no 4 pp 311ndash314 2012

[26] G-D Zhou and T-H Yi ldquoRecent developments on wirelesssensor networks technology for bridge health monitoringrdquoMathematical Problems in Engineering vol 2013 Article ID947867 33 pages 2013

[27] D Ruiz J Urena J C Garcıa C Perez J M Villadangos and EGarcıa ldquoEfficient trilateration algorithm using time differencesof arrivalrdquo Sensors and Actuators A Physical vol 193 pp 220ndash232 2013

[28] E Doukhnitch M Salamah and E Ozen ldquoAn efficientapproach for trilateration in 3D positioningrdquo Computer Com-munications vol 31 no 17 pp 4124ndash4129 2008

[29] B Sadoun and O Al-Bayari ldquoLocation based services usinggeographical information systemsrdquoComputer Communicationsvol 30 no 16 pp 3154ndash3160 2007

[30] D EManolakis ldquoEfficient solution and performance analysis of3-D position estimation by trilaterationrdquo IEEE Transactions onAerospace and Electronic Systems vol 32 no 4 pp 1239ndash12481996

[31] K Kaemarungsi and P Krishnamurthy ldquoAnalysis of WLANrsquosreceived signal strength indication for indoor location finger-printingrdquo Pervasive and Mobile Computing vol 8 no 2 pp292ndash316 2012

[32] S Thirumuruganathan A Detailed Introduction to K-NearestNeighbor (KNN) Algorithm 2010

[33] Q-G Cao K Li Y-J Liu Q-H Sun and J Zhang ldquoRiskmanagement andworkersrsquo safety behavior control in coalminerdquoSafety Science vol 50 no 4 pp 909ndash913 2012

[34] T-H Yi H-N Li and H-M Sun ldquoMulti-stage structuraldamage diagnosis method based on ldquoenergy-damagerdquo theoryrdquoSmart Structures and Systems vol 12 no 3-4 pp 345ndash361 2013

[35] S-F Zhao and L-C Chen ldquoThe application of the integratedindicators based on BP neural network in colliery equipmentsafety monitoringrdquo in Proceedings of the International Confer-ence on E-Product E-Service and E-Entertainment (ICEEE rsquo10)pp 1ndash4 Henan China November 2010

[36] P Lin X L Liu Y Hu W Xu and Q Li ldquoDeformation andstability analysis of Xiluodu arch dam under stress-seepagecoupling conditionrdquo Chinese Journal of Rock Mechanics andEngineering vol 32 no 6 pp 1137ndash1144 2013

[37] P Lin S Z Kang Q B Li and R Wang ldquoEvaluation ofrock mass quality and stability of Xiluodu arch dam underconstruction phaserdquo Chinese Journal of Rock Mechanics andEngineering vol 31 no 10 pp 2042ndash2052 2012

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of


Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of


Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of


Mathematical PhysicsAdvances in

Complex AnalysisJournal of


OptimizationJournal of


CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of


Operations ResearchAdvances in

Journal of


Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Algebra

Discrete Dynamics in Nature and Society



Decision SciencesAdvances in

Discrete MathematicsJournal of


Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of


of arrival (TDOA) and angle of arrival (AOA) GPS is well-known to work independently (defined as a device that doesnot require any installation of technology on a constructionsite other than a device on the resource to position it)for tracking service [19] RFID technology [20] enables aseamless link between any physical tagged entity and thebusiness information infrastructure providing lightweightcomputational and communication capabilities Currentlythe RFID technology that is used for personnel localizationis actually only an attendance recording system rather thanthe real location tracking system Using it not the exact pointof the target but only an area with bigger scope than thetarget can be knownThis does not satisfy the requirement ofreal-time precision positioning When the accidents happenin the tunnels it will be very difficult to rescue the trappedworkers because of low positioning accuracyWireless sensornetworks (WSN) have attracted more and more researchinterest in tunnel applications for their advantages of self-organization low cost and high reliability [24 26] Wirelesssensor networks for location tracking would allow for a widedeployment of sensors across construction sites and as aconsequence a chance for ubiquitous computing capable ofimplementing even complex applications such as integratedproject monitoring to identify the real state of constructionsite execution

With the development of IEEE80211 technology WiFispreads all over the world Its coverage becomes wider andwider Although WiFi is not designed for positioning thesignal that access point (AP) or station regularly sends con-tains the information of RSS which provides the possibilityfor locating mobile station The academia and industry paygreat attentions to applyingWiFi technology to locate pointsRFID has a history of demonstrated ability and marketdominance yet it also has a key disadvantagemdashthe factthat it is nowadays populated with proprietary solutionsincluding expensive readers WiFi-based real-time locatingsystem (RTLS) has recently become just such an opportunitywith RTLS functions being handled by specialized softwareCompared with the existing positioning technology such asGPS cellular localization RFID and ZigBee the positioningbased onWiFi has the following advantages [11] (1) can workon different occasions such as indoor and outdoor providingthe possibility of the ubiquitous positioning (2) only dependson the existing WiFi network does not need to make anychanges and is of low cost which means the existing ITinfrastructure can be reused (3) effect of non-line-of-sight(NLOS) on WiFi signal is small even in the situations wherethere are obstacles With the advances in data mining withbig data wireless communication technologies and smartmobile devices a real-time LBS which provide personalizedservices based on usersrsquo location information and the grow-ing accumulation of industry knowledge have been widelyapplied in military transportation logistics constructionsite and so forth Realization of environment monitoringand worker localization in underground constructions playsan important role in construction and workerrsquos safety Inorder to safeguard the workers firstly need to know wherethey are then we can carry out effective implementation of

LBS that provides rich and extensible services especially afterthe catastrophic accidents and so the rescue team can reachthe disaster scene accurately and carry out the rescue worktimely by acquiring accurate location information Thus insome industry like underground mining installing localiza-tion system is enforced by Chinese Governmentrsquos law Inhydropower industry leading corporations likeThree GorgesCorporation not only actively adopt sensory technologies inmonitoring health of hydraulic structural engineering [7 911] which makes a great contribution to cracking control ofmass concrete [9] but also pay more and more attention toLBS system to protect workersrsquo safety and various services

In this study the positioning system based onWiFi usingRSS is introduced The LBS system provides data based onthe location of the mobile client and can be segmentedinto ldquopushrdquo and ldquopullrdquo models The risk management is alsoproposed This paper studies the feasibility of an LBS systemto provide workersrsquo safety and protection and various servicesin tunnels of high arch dam site The proposed system wastested and deployed inmain tunnels of Xiluodu arch dam site

2 Methodology and Model

21 Position Calculation Method Currently the methods ofmost positioning systems based on WiFi using RSS aredivided into two main categories trilateration algorithm [2728] and fingerprinting [25] technique Trilateration algorithmestimates the target position by measuring the distancebetween target and at least three known reference pointswhile fingerprinting technique gets the target location bymatching the fingerprint information which is the character-istic of signals

211 Trilateration Algorithm Trilateration [27 28] is amethod of calculating the coordinates based on geometrywhich can be trivially expressed as the problem of findingthe intersection of three spheres that involves a systemof quadratic equations [28] There are many algebraic andnumerical methods to solve this problem in 2D [29] andin 3D positioning [28 30] The precondition of a simplifiedgeometric algorithm is that estimated distances from a nodeto at least three of the anchor nodes are known This methodis utilized in the three anchor nodes as the intersection ofa circle centered at the position of the unknown node asshown in Figure 1 The coordinates of the three anchor nodes1198601 1198602 and 119860

3is known in advance as (119909

1198861 1199101198861) (1199091198862 1199101198862)

and (1199091198863 1199101198863) and the distances from these three anchor

nodes to node 119860 are 1198891198861 1198891198862 and 119889

1198863 If the coordinates of

node 119860 (119909 119910) are unknown then there will be the followingformulae [29]

radic(119909 minus 1199091198861)2

+ (119910 minus 1199101198861)2

= 1198891198861

radic(119909 minus 1199091198862)2

+ (119910 minus 1199101198862)2

= 1198891198862

radic(119909 minus 1199091198863)2

+ (119910 minus 1199101198863)2

= 1198891198863

(1)


A3

A1

A2

A

y

x

z



[119909

119910] = [

2 (1199091198861minus 1199091198863) 2 (119910

1198861minus 1199101198863)

2 (1199091198862minus 1199091198863) 2 (119910

1198862minus 1199101198863)]

minus1

times [

[

1199092

1198861minus 1199092

1198863+ 1199102

1198861minus 1199102

1198863+ 1198892

1198861minus 1198892

1198863

1199092

1198862minus 1199092

1198863+ 1199102

1198862minus 1199102

1198863+ 1198892

1198862minus 1198892

1198863

]

]

(2)









Off-line


Estimated location





1198712

119896= 1198751 1198752 119875

119896 (3)


1198861119870

= 1198861 119886

119896 (4)

which satisfies

119889 (119910 minus 1198861) le 119889 (119910 119886

1) (5)

where 119886119894

isin 1198861119870

119886119895

isin 1198861119896




119909 =1

119896

119870

sum

119894=1

119875119894 119875119894isin 1198711119896 (6)


10038161003816100381610038161003816119910119895minus 119886119894119895

10038161003816100381610038161003816gt 2119894119895 (7)







119886 =

119899

sum

119894=1

119883119894119882119894 (8)

Output

Output layer

Hiddenlayer

Inputlayer

I1

I2

In



119910 = 119891 (119886 minus 120579) (9)


119882119895ℎ(119905 + 1) = 119882

119895ℎ(119905) + 120578

120597119864

120597119881119895ℎ

+ 119886 lfloor119882119895ℎ(119905) minus 119882


119882ℎ119905(119905 + 1) = 119882

ℎ119905(119905) + 120578

120597119864

120597119881ℎ119905

+ 119886 [119882ℎ119905(119905) minus 119882

ℎ119905(119905 minus 1)]

(10)










Client (web mobile)

Location server

WiFi AP

Smart phone

Location app server

GIS server

Fingerprint server

Database server












Get

GetGet

Result

Result

Result


Monitoring zone



Nine modules


Worker information




























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra










A3

A1

A2

A

y

x

z



[119909

119910] = [

2 (1199091198861minus 1199091198863) 2 (119910

1198861minus 1199101198863)

2 (1199091198862minus 1199091198863) 2 (119910

1198862minus 1199101198863)]

minus1

times [

[

1199092

1198861minus 1199092

1198863+ 1199102

1198861minus 1199102

1198863+ 1198892

1198861minus 1198892

1198863

1199092

1198862minus 1199092

1198863+ 1199102

1198862minus 1199102

1198863+ 1198892

1198862minus 1198892

1198863

]

]

(2)









Off-line


Estimated location





1198712

119896= 1198751 1198752 119875

119896 (3)


1198861119870

= 1198861 119886

119896 (4)

which satisfies

119889 (119910 minus 1198861) le 119889 (119910 119886

1) (5)

where 119886119894

isin 1198861119870

119886119895

isin 1198861119896




119909 =1

119896

119870

sum

119894=1

119875119894 119875119894isin 1198711119896 (6)


10038161003816100381610038161003816119910119895minus 119886119894119895

10038161003816100381610038161003816gt 2119894119895 (7)







119886 =

119899

sum

119894=1

119883119894119882119894 (8)

Output

Output layer

Hiddenlayer

Inputlayer

I1

I2

In



119910 = 119891 (119886 minus 120579) (9)


119882119895ℎ(119905 + 1) = 119882

119895ℎ(119905) + 120578

120597119864

120597119881119895ℎ

+ 119886 lfloor119882119895ℎ(119905) minus 119882


119882ℎ119905(119905 + 1) = 119882

ℎ119905(119905) + 120578

120597119864

120597119881ℎ119905

+ 119886 [119882ℎ119905(119905) minus 119882

ℎ119905(119905 minus 1)]

(10)










Client (web mobile)

Location server

WiFi AP

Smart phone

Location app server

GIS server

Fingerprint server

Database server












Get

GetGet

Result

Result

Result


Monitoring zone



Nine modules


Worker information




























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra











119909 =1

119896

119870

sum

119894=1

119875119894 119875119894isin 1198711119896 (6)


10038161003816100381610038161003816119910119895minus 119886119894119895

10038161003816100381610038161003816gt 2119894119895 (7)







119886 =

119899

sum

119894=1

119883119894119882119894 (8)

Output

Output layer

Hiddenlayer

Inputlayer

I1

I2

In



119910 = 119891 (119886 minus 120579) (9)


119882119895ℎ(119905 + 1) = 119882

119895ℎ(119905) + 120578

120597119864

120597119881119895ℎ

+ 119886 lfloor119882119895ℎ(119905) minus 119882


119882ℎ119905(119905 + 1) = 119882

ℎ119905(119905) + 120578

120597119864

120597119881ℎ119905

+ 119886 [119882ℎ119905(119905) minus 119882

ℎ119905(119905 minus 1)]

(10)










Client (web mobile)

Location server

WiFi AP

Smart phone

Location app server

GIS server

Fingerprint server

Database server












Get

GetGet

Result

Result

Result


Monitoring zone



Nine modules


Worker information




























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra










Client (web mobile)

Location server

WiFi AP

Smart phone

Location app server

GIS server

Fingerprint server

Database server












Get

GetGet

Result

Result

Result


Monitoring zone



Nine modules


Worker information




























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra











Get

GetGet

Result

Result

Result


Monitoring zone



Nine modules


Worker information




























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra






























Drainage tunnel


Grouting tunnel

Elevator shaft

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Monitoring tunnel

Ventilation shaft

Platform


669 365 11233677 357 11233674 360 11233673 352 11233365 483 11233376 486 15561376 486 15561453 491 15561453 491 15561

672 362 16069670 349 16069672 362 16079672 362 16079672 362 16079672 362 16079670 349 16079670 349 16079670 349 16079


5656427

5656428

5656429

5656719

5656720

5656721

5656722

5656723


5656898

5656899

5656900

5656901

5656902

5656903

5656904

5656905

5656906







Risk (119863) = 119871 times 119864 times 119862 (11)



1 1198832 119883

119899(where 119899 gt 200) And

11988211198822 119882







Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra










Initialization

Input training data

Computing

Get error value

E lt limit value

End

Change weight value


05

1015202530

Valu

e

Risk

2013

10

1

2013

10

3

2013

10

5

2013

10

7

2013

10

9

2013

10

11

2013

10

13

2013

10

15

2013

10

17

2013

10

19

2013

10

21

2013

10

23

2013

10

25

2013

10

27

2013

10

29


5 Conclusions







Acknowledgments


References















































Volume 2014




Journal of











Journal of


Function Spaces






Algebra











































Volume 2014




Journal of











Journal of


Function Spaces






Algebra









research article a real-time location-based services

Documents