Research Article Optimal Network QoS over the...

Research ArticleOptimal Network QoS over the Internet ofVehicles for E-Health Applications

Di Lin1 Yuanzhe Yao1 Fabrice Labeau2 Yu Tang1 and Athanasios V Vasilakos3

1School of Information and Software Engineering University of Electronic Science and Technology of China Chengdu 610000 China2Electrical and Computer Engineering Department McGill University Montreal QC Canada H3A 0G43Department of Computer Science Electrical and Space Engineering Lulea University of Technology 97187 Lulea Sweden

Correspondence should be addressed to Di Lin dilin2mailmcgillca

Received 11 November 2015 Accepted 26 January 2016

Academic Editor Lingjie Duan

Copyright copy 2016 Di Lin et al This 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

Wireless technologies are pervasive to support ubiquitous healthcare applications However a critical issue of using wirelesscommunications under a healthcare scenario is the electromagnetic interference (EMI) caused by RF transmission and a highlevel of EMI may lead to a critical malfunction of medical sensors In consideration of EMI on medical sensors we study theoptimization of quality of service (QoS) within the whole Internet of vehicles for E-health and propose a novel model to optimizethe QoS by allocating the transmit power of each user Our results show that the optimal power control policy depends on theobjective of optimization problems a greedy policy is optimal to maximize the summation of QoS of each user whereas a fairpolicy is optimal to maximize the product of QoS of each user Algorithms are taken to derive the optimal policies and numericalresults of optimizing QoS are presented for both objectives and QoS constraints

1 Introduction

Recent developments in cellular networks have enabledubiquitous applications of E-health with the aid of medicalsensors However RF transmission in cellular networks canresult in electromagnetic interference (EMI) to medical sen-sors and a high level of interference can cause malfunction ofmedical sensors and even injure patients [1 2]Thus the con-trol of EMI (eg through power control) is a critical issue toE-health and should be investigated under the environmentof mobile hospital which is defined as Internet of vehicles forE-health applications throughout this paper Alternatively weuse the terms of mobile hospital and Internet of vehicles forE-health applications

11 Motivation and Novelty EMI can cause the malfunctionof medical sensors The network analysis in considerationof EMI control (eg through power control) is a criticalissue to the E-health networks However the algorithmsof network analysis in regular wireless networks [3 4]cannot be employed in E-health networks since the former

does not take into account the impact of EMI on networkperformance For example in E-health network the wirelessusers have to reduce their transmit power to control theEMI impact and may not achieve the same level of systemcapacity estimated by the algorithms used in a regularwireless network

A large number of works are related to the application ofwireless networks in order to support health service [1 2 5]Phunchongharn et al in [1 2] address the issue of EMI underthe scenario of wireless local area networks (WLAN) forE-health applications within a hospital but the technologyof WLAN cannot be applied in our scenario in which anInternet of vehicles covers a large-scaled area (eg a city or adistrict) Shen et al in [5] present the possibilities of employ-ing wireless technologies in a medical environment andadjust the level of power and rate with the channel conditionsof users However the authors do not take the potential EMIimpact into account In such a scenario a wireless user who isclose to a medical sensor could be allowed to transmit data ata high level of power when the userrsquos communication channelis in good condition [6] However the RF transmission at

Hindawi Publishing CorporationMobile Information SystemsVolume 2016 Article ID 5140486 11 pageshttpdxdoiorg10115520165140486

2 Mobile Information Systems

a high level of power may lead to the malfunction of medicalsensors Such an improper power allocation by the above-mentioned algorithms may influence the function of EMI-sensitive medical sensors so these algorithms cannot beemployed under the scenario of mobile hospital Also theabove-mentioned algorithms are designed to optimize theindividual objective of each wireless user instead of opti-mizing a network-level objective (eg the quality of networkservice) The importance of scheduling wireless transmissionunder a mobile hospital scenario and the lack of efficientalgorithms for optimizing network-level objective motivate usto study how wireless users can control their transmit powerto achieve certain goals such as maximizing the quality of net-work service while ensuring the acceptable level of EMI onmed-ical sensors over Internet of vehicles for E-health applications

12 Main Contributions In this paper we present the issueof scheduling wireless transmission under the environmentof mobile hospital The objective of this paper is to optimizecertain goals (eg the quality of service) at the network levelinstead of at the user level In this paper we address theproblem of optimizing the quality of network service in amobile hospital environment and propose the algorithm ofpower control to achieve the optimal network service To thebest of our knowledge this is the first work which addressesthe algorithm of power control to achieve a network-levelgoal under a wireless network for E-health applications Theprimary contributions of this paper are composed of the fol-lowing issues (i) establishing a problem of optimizing powercontrol to achieve the globally optimal quality of service atthe network level in view of EMI on medical sensors (ii)analyzing the transmit power and quality of service of eachuser at the optimal solution to the proposed problem

2 Related Work of EMI on Medical Sensors

In hospital scenarios the research on EMI begins with thestudy of immunity of medical equipment to mobile phonesTan and Hinberg in [7] address that a few pieces of medicalequipment are sensitively influenced by the EMI frommobilephones and the typical pieces of equipment include infusionpumps ventilators and ECGmonitors Also an EMI suscep-tibility test was conducted by the Medicines and HealthcareProducts Regulatory Agency (MHRA) of UK [8] This testfocuses on investigating the EMI of mobile phones within apersonal communication network and its results show thatthe EMI-sensitive medical equipment include respiratorsdefibrillators and external pacemakers Trigano et al in [9]and Calcagnini et al in [10] present how the EMI fromGSM mobile phones influences pacemakers and infusionpumps respectively The results show that the EMI of mobilephones can lead to the malfunction of both pacemakers andinfusion pumps With the wide use of 3rd-generation (3G)telecommunication systems all over the world the authors in[11 12] study how EMI impacts medical pieces of equipmentwithin the 3G bands Based on the aforementioned researchthe International Electrotechnical Committee (IEC) in 2007publishes the EN60601-1-2 standard and this standard rec-ommends the level of EMI immunity as 10Vm and 3Vm

for non-life-supporting equipment (eg defibrillators) andlife-supporting equipment (eg blood pressuremonitors andinfusion pumps) respectively In view of the advances ofelectromagnetic compatibility (EMC) technologies Singa-pore and the UK governments relax the EMI restrictionwhich is recommended by EN60601-1-2 standard andmobilephones are allowed to be uses in certain areas of hospitals[13] In consideration of the most advanced EMC of medicalequipment Tang et al in [14] address an EMI test whichinvestigates the EMI fromGSM900 PCS1800 and 3Gmobilecommunication systems This test shows that EMI-sensitiveequipment includes ECG monitors audio evoked potentialsystems radiographic systems and ultrasonic fetal heartdetectors [14] From the study results of previous literaturewe can reach a conclusion syringe pumps ECG monitorsfetal monitors respirators anesthesia machines externalpacemakers infusion pumps and defibrillators are sensitiveto the EMI of mobile phones [15]

The other studies focus on the EMI from devices whichhave access to a wireless local area network (WLAN) anda typical WLAN usually works around the frequency of24GHz which is different from the working band of mobilephones The amount of EMI on medical equipment is par-tially determined by frequency bands and thus the researchon EMI under the scenario of healthcare monitoring withina WLAN appears Krishnamoorthy et al in [16] carry out ameasurement of EMI caused by doctor devices on pieces ofmedical equipment located in three hospitals and the doctordevices work within a 24GHz-band WLAN The measure-ments show that the highest level of EMI is 0552Vmwhich is in the acceptable EMI range recommended by theEN60601-1-2 standard However the test in [16] did not takeinto account the quality of service (QoS) of data whichare transmitted between patient devices and doctor devicesThe strategies of restricting the use of mobile phones suchas switching off mobile phones are not applicable for thescenario of a wireless healthcare monitoring system [17] Inwireless healthcare monitoring systems doctors and patientsneed to employwireless devices to transmit data and establishcommunication and the restriction on the level of transmitpower may cause a lower level of QoS of data transmissionwhich would lead to the loss ofmedical dataThus in wirelesshealthcare monitoring systems the restriction of transmitpower and QoS requirements are usually against each otherAdditionally the simultaneous transmission of data amongmultiple patient devices and doctor devices would cause ahigh level of EMI to medical equipment [1] Phunchongharnet al in [1] investigate the EMI in hospital environments byconsidering the QoS of patient devices and doctor devicesand they conclude that the level of EMI on most of themedical equipment is unacceptable when the transmit powerof a device is above 10mW within a WLAN

All the above-mentioned research does not consider thevehicular scenarios for healthcare applications which areinteresting to this paper and thus the medical sensors inthe test may not be vehicle-mounted and wearable medicalsensors In Section 31 we address a detailed experimentwhich includes the test of EMI impact on types of vehicle-mounted and wearable medical sensors

Mobile Information Systems 3

3 Mobile Hospital Environment

In a mobile hospital environment vehicles for E-healthapplications are mounted with a few medical devices andthese devices can assist doctors in monitoring the patientsrsquocondition (to the best of our knowledge quite a few Internetof medical vehicles examples have already been employed inUSA to provide mobile health service A few real examplesof service providers include Mobile Specialty Vehicles andFarber Specialty Vehicles) Under the following scenariosphysicians nurses and the patientrsquos relatives may employmobile phones (1) Physicians and nurses staff need to keepreporting the patientsrsquo conditions over phone to a doctoror healthcare staff in the medical center Once the patientsarrive at the medical center the staff can arrange the medicalactions on this patient (2) Relatives of patients may need tocontact their families or friends by phone about the importantinformation for example the variation of clinical situationsHowever the nearby medical devices might be impacted bythe EMI which is produced by the use of mobile phones [18]EMI is defined as the disturbance of electrical circuits causedby electromagnetic radiation from an external source [19]and EMI could lead to loss of data during the transmission

In the following we address an experiment of EMI effectson typical medical devices This experiment is originallyproposed in [20] and we present it here for completenessThen we establish the model of EMI impact which is aconstraint of our network-capacity optimization problemdetailed in Section 4

31 Experiment of Investigating EMI Impact In the followingexperiment we investigate how the EMI of mobile phonesimpacts a few typical vehicle-mounted medical devices Forcompleteness we choose the mobile phones with quite a fewtypical technologies including CDMA2000 TD-LTE andGSM-9001800 We carry out this experiment in an anechoicchamber which can exclude the EMI impact caused by theother RF sources such as from external communicationsystems

The results of experiment show that the EMI frommobilephones could degrade the performance of medical deviceswhen these devices keep a distance of 2m from the mobilephones Typical types of degradation in the experimentinclude (I) errors occurring in the ultrasound X-ray and CTimages (II) artifact in ECG and EEG signals (III) the mal-function of ventilators syringe pumps and infusion pumpsand (IV) irregular operating modes of external pacemakersThe above-mentioned results are in line with the publicationsof [21ndash24]

32 Model of EMI Impact Under the scenario of E-healthapplications a vehicle is mounted with medical devicesand these devices are either life-support or non-life-support(illustrated in Figure 1) These medical devices first collectmedical data and then send these data to physicians or doc-tors in order to takemedical actions on a specific patient oncethis patient arrives Also physicians or healthcare staff on thevehicle need to report the latest conditions of a patient todoctors ensuring that the doctors could acquire the updated

EMI to medical sensors

Vehicle for E-health

Mobile station

Holter

Ultrasonographysensor

Blood pressuresensor

Holter

Medical datatransmission

EMI impact

Hospital

Reporting patientcondition (talking

over phone)




Figure 1 The figure illustrates the Internet of vehicles for E-healthapplications

patientrsquos conditions However nearby medical devices mightbe impacted by the EMI from the use of mobile phones Alsolife-supportmedical devices aremore sensitive to EMI impactthan non-life-support devices since the former containselectronic components which are EMI-sensitive Typical life-support medical devices are Ultrasonograph devices and soforth and typical non-life-support medical devices are bloodpressure devices and holters and so forth

Different medical devices can allow different levels oftransmit power to avoid the malfunction of these devicesTo the best of our knowledge Phunchongharn et al in [1]firstly present the models of EMI on medical devices andinvestigate the highest allowable level of transmit power tomeet the EMI constraint The maximal allowable transmitpower of a mobile phone needs to satisfy (1) and (2) fornon-life-support medical devices and life-support medicaldevices respectively [1]

sum119895isin119878

1205791radic119875119895119863119895 (119909)

le 119864NL (119909) for 119909 isin 1198781 (1)

sum119895isin119878

1205792radic119875119895119863119895 (119910)

le 119864L (119910) for 119910 isin 1198782 (2)


where119864NL(119909) and119864L(119910) are the allowable EMI levels of a non-life-support device 119909 and a piece of life-support equipment119910 respectively 119875119895 is the transmit power of a mobile user119895 119863119895(119909) represents the distance between a non-life-supportdevice 119909 and user 119895 119863119895(119910) represents the distance betweena life-support device 119910 and user 119895 1205791 and 1205792 represent twoconstants the values of which are recommended by IEC60601-1-2 as 7 and 23 respectively [1] 119878 represents the set ofmobile users under the scenario of Internet of vehicles 1198781 isthe set of non-life-support devices while 1198782 is the set of life-support devices

Definition 1 The maximal possible transmit power of user 119894(ie 119875119894) to satisfy the constraint of EMI on medical devicescan be computed from (1) and (2) Also we define theminimum of wireless usersrsquo allowable transmit power 119875119894 thatis 119875max = min119894119875119894 as the maximal effective transmit power(METP) under a mobile hospital scenario (for the detailedprocess of computing METP refer to [20])

METP 119875max will be employed to establish the optimiza-tion problem in Section 4 Each of the mobile users needsto keep his or her transmit power below METP ensuring anallowable level of EMI under the mobile hospital scenario (ina mobile health scenario the wireless users may move to anyposition around a piece ofmedical equipment andwe assumethat each of the users has the same probability to appear at aspecific position So the users close to the same equipmentequally share a bounded transmit powerThe investigation ofdifferentiated bounds of transmit power for different wirelessusers (physicians nurses patients and their relatives) wouldbe an interesting topic but it is out of the scope of this paper)

4 Optimization of QoS

In this section we firstly summarize the mathematical for-mulation of QoS Consider an Internet of vehicles with 119873wireless users For information user 119894 Signal-to-Noise Ratio(SNR) is defined as

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868 (3)

where 119875119894 denotes the transmit power by user 119894 ℎ119894 denotes thechannel condition between user 119894 and base station 119868 denotesthe power of additive white Gaussian noise

In the following wewill employ SNR as themetric of QoSand discuss how to optimize the QoS within a network ofmobile hospital Given the metric of QoS of each user as (3)we consider the problem of optimizing QoS within a networkwith two objectives as

1198761 = sum119894

SNR119894 (4)

1198762 = prod119894

SNR119894 (5)

The problem objective (5) represents the total throughputmaximization with fixed coding Suppose we have alreadychosen a specific scheme of coding at the symbol level

that is we have spreading gain and power at our controlThis problem is formulated in [6] As shown there the totalthroughput is always proportional to the level of SINR wherethe proportion is determined by the coding scheme Conse-quently optimizing the total throughput equals maximizingthe summation of SINRThe problem objective (6) representsthe total throughput maximization with flexible codingSuppose we relaxed the limitation of selecting and fixing aspecific scheme of coding at the symbol level This problemis formulated in [6] As shown there the total throughputis always proportional to the level of log function of SINRwhere the proportion is determined by the coding schemeConsequently optimizing the total throughput equals max-imizing the product of SINR The problem of optimizingtotal throughput either with fixed coding scheme or withflexible coding scheme is a critical problem in network andthis motivates us to investigate the optimization of systemthroughput in the scenario of wireless network of E-healthwith the design of objectives of (5) and (6)

In the following we formulate the optimization problemswith 1198761 and 1198762 as

maxP

sum119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(6)

maxP

prod119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(7)

where P denotes the optimal transmit power of each user119875max denotes the maximal possible transmit power among allof the users SNRmin denotes the minimal level of SNR whichis acceptable to all of the users 119875119878 denotes the total receivedpower constraint It is due to the fact that the total receivedpower from the data users (for E-health applications) isan interference to other classes of users (not for E-healthapplications)

With the setting of 119909119894 = 119875119894ℎ119894 we can transform theoptimization problems as

maxP

sum119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894minus 119909119894 + 119868 times SNRmin +sum

119895 =119894

119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(8)


maxP

prod119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894

minus 119909119894 + 119868 times SNRmin +sum119895 =119894

119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(9)

where 119909119894 denotes the level of power at the receiving end ofuser 119894

The Lagrange formulation of optimization problem withobjective 119876 (119876 = 1198761 1198762) can be denoted as

119871 = 119876 minussum119894

120579119894 (119909119894ℎ119894minus 119875max)

minussum119894

120582119894(minus119909119894 + 119868 times SNRmin +sum119895 =119894

119909119895)

minus 120583(sum119894

119909119894 minus 119875119878)

= 119876 minussum119894

(120579119894ℎ119894+ 120583 minus 120582119894 + SNRminsum

119895 =119894

120582119895)119909119894

+ Const

(10)

where 120583 120582119894 and 120579119894 are the parameters of Lagrange formula-tion

Then the first-order derivative of 119871 is

120597119871120597119909119894

= 0 997888rarr

120579119894ℎ119894+ 120583 minus 120582119894 + SNRminsum

119895 =119894

120582119895 =120597119876120597119909119894

(11)

Note that 119875119878 is on all the users so the value of 120583 iscommon to all the usersWe can classify all119873 users into threedisjoint groups according to the binding constraints that isdepending on the values of 120582119894 andor 120579119894

Definition 2 Group 1 1198661 = 119894 | 120582119894 gt 0 which is equallydefined as 1198661 = 119894 | SNR119894 = SNRmin Group 2 1198662 = 119894 |120582119894 = 0 120579119894 = 0 which is equally defined as 1198662 = 119894 | SNR119894 gtSNRmin 119875119894 lt 119875max Group 3 1198663 = 119894 | 120582119894 = 0 120579119894 gt 0 whichis equally defined as 1198663 = 119894 | SNR119894 gt SNRmin 119875119894 = 119875max

Remark 3 In consideration of EMI andQoS requirements allof the users cannot violate the constraints of EMI and QoSGroup 1 represents the set of users who are at the verge ofviolating QoS constraint Group 3 represents the set of userswho are at the verge of violating EMI constraint Group 2represents the set of users who are not at the verge of violatingany constraint

5 Algorithms of SolvingOptimization Problems

In the following we first derive the general structure of theoptimal solution to the problems with objectives 1198761 and 1198762respectivelyThen we address the specific algorithms to solvethe optimization problems

51 Solutions to the Optimal Problem with Objective 1198761

Lemma 4 One has sign(1205971198761120597119909119894 minus1205971198761120597119909119895) = sign(119909119894 minus119909119895)where sign(119909) represents the sign of 119909

Proof Given

1205971198761120597119909119894

= 1119868 + sum119873119896=1119896 =119894 119909119896

minus119873

sum119898=1119898 =119896

119909119898(119868 + sum119873119896=1119896 =119898 119909119896)

2 (12)

we have

1205971198761120597119909119894

minus 1205971198761120597119909119895

= (119868 +119873

sum119896=1119896 =119894

119909119896)

sdot

1

(119868 + sum119873119896=1119896 =119894 119909119896)2minus 1

(119868 + sum119873119896=1119896 =119895 119909119896)2

(13)

If 119909119894 le 119909119895 we have 1205971198761120597119909119894 minus 1205971198761120597119909119895 le 0 otherwise1205971198761120597119909119894 minus 1205971198761120597119909119895 lt 0 So sign(1205971198761120597119909119894 minus 1205971198761120597119909119895) =sign(119909119894 minus 119909119895)

Lemma 5 For an optimum solution the number of users ingroup 1198662 is at most one

Proof Suppose that we can find two users 119894 and 119895 in group1198662 which can lead to the optimum xlowast then 1205971198761120597119909lowast119894 =1205971198761120597119909

lowast119895 = 120583 + sum119895notin1198662 120582119895 given Definition 1

Let x120598 = 119909lowast1 119909lowast119894 +120598 119909

lowast119895 minus120598 119909119873 then1198761(x

120598)minus1198761(xlowast) = int

120598

0(12059711987611205971205980)1198891205980 gt 0 since 12059711987611205971205980 = (1205971198761120597119909119894 minus

1205971198761120597119909119895)|x=x1205980

gt 0 for any 0 le 1205980 le 120598 1198761(x120598) gt 1198761(xlowast) is incontradiction with the fact that xlowast is the optimum of1198761

Theorem 6 When the number of users in 1198662 is zero then wecan restrict our attention of an optimal solution to the followingcases without losing optimality any user 119895 isin 1198663 has a betterchannel condition than any user 119894 isin 1198661 if 119875119894 lt 119875max that isℎ119894 lt ℎ119895 for 119895 isin 1198663 and 119894 isin 1198661 with 119875119894 lt 119875max

Proof When the number of users in1198662 is zero then any userbelongs to1198661 or1198663 Suppose user 119895 isin 1198663 and 119894 isin 1198661 with119875119894 lt119875max and xlowast = 119909lowast1 119909

lowast119894 119909

lowast119895 119909

lowast119873 is the optimum

One has 1205971198761120597119909lowast119894 minus 1205971198761120597119909

lowast119895 = minus120579119895ℎ119895 minus (1 + SNRmin)120582119894 lt 0

By Lemma 4 we have 119909lowast119894 lt 119909lowast119895

Let xlowastlowast = 119909lowast1 119909lowast119895 119909

lowast119894 119909

lowast119873 that is we

exchange the position of 119894th and 119895th items of xlowast Assumeℎ119894 ge ℎ119895 Since 119909

lowast119894 lt 119909lowast119895 and 1198761(x

lowastlowast) = 1198761(xlowast) we knowthat xlowastlowast is also an optimal solution to problem (6) Howeverxlowastlowast has a user 119895 who belongs to 1198662 Thus the assumption ofℎ119894 ge ℎ119895 leads to a contradiction


Theorem7 When the number of users in1198662 is 1 the userswithworse channel condition than the users in1198662must belong to1198661that is any user 119894 with ℎ119894 lt ℎ119895 given 119895 isin 1198662 satisfies 119894 isin 1198661

Proof Suppose that xlowast = 119909lowast1 119909lowast119894 119909

lowast119895 119909

lowast119873 is the

optimum Consider users 119895 isin 1198663 and 119894 isin 1198662 One has1205971198761120597119909

lowast119894 minus 1205971198761120597119909

lowast119895 = minus120579119895ℎ119895 lt 0 By Lemma 4 we have

119909lowast119894 lt 119909lowast119895

Assume ℎ119894 ge ℎ119895 Following the same way as that used inthe proof ofTheorem 6 by exchanging the position of 119894th and119895th items of xlowast we can find an optimum xlowastlowast with two userswho belong to 1198662 Thus the assumption of ℎ119894 ge ℎ119895 leads to acontradiction

All of users in1198663 have a better channel condition than theuser in 1198662 Thus if a user 119896 with ℎ119896 lt ℎ119894 given 119894 isin 1198662 thenuser 119896 belongs to 1198661

Theorem 8 When the number of users in 1198662 is 1 we canrestrict our attention of an optimal solution to the followingcases without losing optimality all the users with better channelcondition than the user in 1198662 belong to 1198663 that is any user 119894with ℎ119894 gt ℎ119895 given 119895 isin 1198662 satisfies 119894 isin 1198663

Proof Refer to Theorem 7

Remark 9 The optimal solution to the problem with 1198761implies a greedy policy the user 119894 with better channelconditions (a higher ℎ119894) is allocated to 119875max (in group1198663) theuser 119894 with worse channel conditions (a lower ℎ119894) is allocatedto a level of power which can only meet the minimal QoS (ingroup 1198661) and the number of users in 1198662 is at most one


Lemma 10 One has sign(1205971198762120597119909119894 minus 1205971198762120597119909119895) = minus sign(119909119894 minus119909119895) where sign(119909) represents the sign of 119909

Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)

If 119909119894 le 119909119895 we have 1205971198762120597119909119894 minus 1205971198762120597119909119895 ge 0 otherwise1205971198762120597119909119894 minus 1205971198762120597119909119895 gt 0 So sign(1205971198762120597119909119894 minus 1205971198762120597119909119895) =minus sign(119909119894 minus 119909119895)

Theorem 11 For an optimum solution if there are users in1198662their level of power at the receiving end should be the same Inother words if there are users 119894 119895 isin 1198662 then 119909119894 = 119909119895

Proof Consider users 119894 119895 isin 1198662 We have 1205971198762120597119909119894 =1205971198762120597119909119895 = 120583 + SNRminsum119896isin119860

1

120582119896 So 119909119894 = 119909119895 by Lemma 10

Theorem 12 Except for the case when all the usersrsquo powers atthe receiving end 119909119894 (119894 = 1 2 119873) are the same if there areusers in 1198661 then these users should be allocated the maximaltransmit power that is 119875max


lowast119895 119909

lowast119873 is the

optimum Consider users 119894 isin 1198661 and 119895 notin 1198661 Assume119875119894 lt 119875max Then we have 1205971198762120597119909119894 minus 1205971198762120597119909119895 lt 0 and thus119909119894 gt 119909119895 by Lemma 10 However from the monotonicity ofSNR equation shown in (3) we have SNR119894 = SNRmin gt SNR119895which is a contradiction Thus we have 119875119894 = 119875max

Theorem 13 Except for the case when all the usersrsquo powers atthe receiving end 119909119894 (119894 = 1 2 119873) are the same there can beat most one user in 1198661 and that user is with the worst channelcondition ℎ119894 In other words the number of users in 1198661 is atmost 1 and ℎ119894 is lowest if 119894 isin 1198661


lowast119895 119909

lowast119873 is the

optimum Assume that there are two users 119894 119895 isin 1198661 with ℎ119894 gtℎ119895 By Theorem 12 119875119894 = 119875119895 = 119875max So we have 119909119894 gt 119909119895 andSNR119894 gt SNR119895 from themonotonicity of SNR equation whichis a contradiction with SNR119894 = SNR119895 = SNRmin So there isat most one user in 1198661

Theorem 14 For an optimum solution if there exists user 119894 in1198663 then 119909119894 should be less than 119909119895 for any user 119895 isin 1198662


lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909

lowast119895 = 120579119895ℎ119895 gt 0 By Lemma 10 we have

119909lowast119894 lt 119909lowast119895

Remark 15 The optimal solution to the problem with 1198762implies a fair policy the user 119894with worse channel conditions(a lower ℎ119894) is allocated to 119875max (in group1198661 or1198663) the user 119894with better channel conditions (a higher ℎ119894) is allocated to thesame level of power at the receiving end 119909119894 which is higherthan 119909119895 of user 119895 who is with lower priority (a lower ℎ119895)

53 Algorithms of Solving the Optimization Problem

Theorem 16 For problem of (6) without losing optimality wecan focus our attention on an optimal power allocation vectorto the following form Plowast = 119875max 119875max 119875119879 119875119879+1 119875119873where 119879 is an integer and 1 le 119879 le 119873 0 le 119875119879 le 119875max SNR119894 =SNRmin for any 119879 lt 119894 le 119873

Proof Refer to Remark 3 andTheorems 6 to 8

FromTheorem 16 an optimal solution to problem (6) canbe found as follows Partition the users into two classes 119879low-priority users and (119873 minus 119879) high-priority users basedon their channel conditions ℎ119894 (119894 = 1 2 119873) (a userin a better channel condition ie a larger value of ℎ119894 isallocated a higher priority) For each of the 119873 + 1 possibleways of partitioning users allocate the power in the followingway Allocate 119875max to all high-priority users while the poweris allocated to all low-priority users just to meet the QoS


Ensure 0 le 119875119894 le 119875max SNR119894 ge SNRminsum119894 119875119894ℎ119894 le 119875119878Initialize P0= 119875max 119875max 0 0 0119879 = 1119906 = 119875maxℎmin (ℎmin = min119894ℎ119894)1198761 = 0while 1 le 119879 le 119873 do

whilemin119894SNR119894 ge SNRmin do119876 = sum119894 SNR119894if 119876 gt 1198761 then1198761 = 119876P119879 = 119906ℎ119879 and update P

end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1

end whileOutput the result of 1198761 and P

Algorithm 1 Algorithm of solving optimization problem (6)

requirement that is SNRmin Optimize the objective withrespect to the variable 119879 subject to the constraints of (6)Evaluate the objective and compare the values of objective foreach possible partitioning rule Pick119879 that yield the optimumby increasing 119906 at a step of Δ119875 The specific algorithm isshown in Algorithm 1

Remark 17 Algorithm 1 solves the problem of (6) within therunning time of 119874(119873119870) where 119870 = lceil119875maxℎminΔ119875rceil andℎmin = min119894ℎ119894

Remark 18 Exploratory search algorithms solve the problemof (6) within the running time of 119874(119870119873) where 119870 =lceil119875maxℎminΔ119875rceil

Theorem 19 For problem of (7) without losing optimality wecan focus our attention on an optimal power allocation vectorto the following form Plowast = 1198751 119875119879 119875max 119875max where119879 is an integer and 1 le 119879 le 119873 0 le 119875119879 le 119875max 1199091 = 1199092 = sdot sdot sdot =119909119879 ge 119909119879+1 = 119875maxℎ119879+1


FromTheorem 19 an optimal solution to problem (7) canbe found as follows Partition the users into two classes 119879low-priority users and (119873 minus 119879) high-priority users based ontheir channel conditions ℎ119894 119894 = 1 2 119873 For each of the119873 + 1 possible ways of partitioning users allocate the powerin the following way Allocate 119875max to all low-priority userswhile the power is allocated among high-priority users toensure that their 119909119894 (119894 = 119879 + 1 119873) is the same (119909119894 =119906 (119894 = 119879 + 1 119873)) and higher than 119875maxℎ119895 (119895 = 1 119879)Optimize the objective with respect to two variables 119879 and119906 subject to all the constraints Evaluate the objective andcompare the values of objective for each possible partitioningrule Pick 119879 and 119906 that yield the optimum by increasing 119906 ata step of Δ119875 The specific algorithm is shown in Algorithm 2

Ensure 0 le 119875119894 le 119875max SNR119894 ge SNRminsum119894 119875119894ℎ119894 le 119875119878InitializeP = 0 0 119875max 119875max119879 = 1119906 = 01198762 = 0while 1 le 119879 le 119873 do

while 119906 le 119875maxℎmin (ℎmin = min119894ℎ119894) do119876 = sum119894 SNR119894if 119876 gt 1198762 then1198762 = 119876P119879 = 119906ℎ119879 and update P

end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1



Remark 20 Algorithm 2 solves the problem of (7) within therunning time of 119874(119873119870) where119870 = lceil119875maxℎminΔ119875rceil


6 Simulation Results

We collect the Internet-of-vehicles data from [26] in whicha transmit-receive pair of mobile users is represented by aconnection of network In the following simulation we set 50nodes in the vehicle network each node using amobile phonewith a probability of 01 Please note that 50 mobile terminalscan produce EMI impact onmedical devices at the same timein a densely populated city Also we set the average distancebetween mobile users as 8 meters Each of the users canmove at a speed of 10ms (36 kmh) in an arbitrary directionWe clarify the model of channel characteristics in Section 51and perform 100000 Matlab-based runs in the experiment toaddress the results The levels of EMI 119864LS or 119864NLS (see (1) and(2)) are normalized to unity in the simulation

61 Characteristics of Channel Models We firstly select a fewtypical empirical channel models for simulation and thesemodels are presented in ITU-R recommendation M1225[25] The ITU-R M1225 model can be applied under thescenarios of mobile hospitals outside the high-rise core ofurban areas in which the height of buildings is almostuniform [25] Mathematically the channel characteristics 119871can be addressed as

119871 = 40 (1 minus 4 times 10minus3Δℎ) log119863 minus 18 logΔℎ + 21 log 119888

+ 80(16)

where 119863[km] refers to the distance between base stationand mobile terminals 119888[MHz] is the frequency of carriers


0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR

20 30 40 5010Size of network

(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)

Figure 2 The figure illustrates the change of SNR with various sizes of networks 119873 (objective of 1198761 versus objective of 1198762) (a) Averagelevel of SNR (b) standard deviation of SNR Green line with ldquoIrdquo represents the problem with objective 1198761 blue line with ldquo998779rdquo represents theproblem with objective 1198762

ℎ[m] refers to the height of antenna at the base station andthe height is measured from an average roof-top level

The authors in [25] characterize each terrestrial environ-ment as a channel impulse response by using tapped-delaylines Specifically the model is composed of quite a few tapsthe time delay is determined by the first tap the average levelof power is determined by the strongest tap Table 1 addressesthe propagation model for all of vehicular experiment casesIn each of these cases Table 1 lists the strength of signalsthe relative time delay and Doppler shift Specifically theprimary parameters of characterizing propagationmodels are

(i) time delay the structure of spread and its probabilitydistribution

(ii) multipath fading characteristics Doppler spectrumin Rician channels versus in Rayleigh channels

62 QoS in the Network of Mobile Hospital In this sectionwe address the average level of SNR in a network of mobilehospital as well as the difference among individual SNRlevels with different objectives of optimizing QoS Alsowe investigate how the parameters of network size 119873 themaximal possible transmit power 119875max and the power ofnoise 119868 can impact the SNR level

We set 119875max = 1 and 119868 = 10 It is observed fromFigure 2 that the optimization problem with objective1198761 canachieve a higher level of average SNR than that with objective1198762 while the latter can achieve a much lower level of SNRdifference among users This is because the optimal solutionto the problem with 1198761 implies a greedy policy the usersin better channel conditions (ie the users with a higherℎ119894) are allocated to 119875max while the users in worse channel

Table 1 Parameters of propagationmodels in ITU-R recommenda-tion M1225 [25]

Tap Relativedelay (ns)

Averagepower (dB)

Dopplerspectrum

1 0 00 Rayleigh2 310 minus10 Rayleigh3 710 minus90 Rayleigh4 1090 minus100 Rayleigh5 1730 minus150 Rayleigh6 2510 minus200 Rayleigh

conditions are allocated to a low level of power which canonly meet the minimal level of QoS However the optimalsolution to the problemwith1198762 implies a fair policy the usersinworse channel conditions (ie the users with a lower ℎ119894) areallocated to 119875max Thus the difference of SNR among userswith objective 1198762 is much lower than that with objective 1198761

We set 119873 = 50 It is observed from Figure 3 that theoptimization solution is influenced by both the parameters of119875max and 119868With the increase of119875max the average level of SNRincreases first and then decreases suggesting that there is athreshold of 119875max beyond which noise has a more significantimpact on SNR than signal does Unsurprisingly with theincrease of 119868 (ie the decrease of 1119868) the value of SNR alsodecreases However the impact of high values of 119868 (low valuesof 1119868) on the SNR decreases significantly as the 119868 increasessuggesting that there is a threshold of noise power 119868 belowwhich additional increase of average level of SNR ceases to beadvisable


0

005

01

015

02

025

Aver

age l

evel

of S

NR

04 06 0802Maximal transmit power Pmax

(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)

Figure 3 The figure illustrates the change of SNR with parameters of 119875max and 119868 (objective of 1198761 versus objective of 1198762) (a) Average levelof SNR versus 119875max (b) average level of SNR versus 119868 Green line with ldquoIrdquo represents for the problem with objective 1198761 blue line with ldquo998779rdquorepresents for the problem with objective 1198762

times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)

Figure 4 The figure illustrates the convergence rate with parameters of network size119873 and 119875max (objective of 1198761 versus objective of 1198762) (a)Convergence rate versus119873 (b) convergence rate versus 119875max Green line with ldquoIrdquo represents for the problem with objective1198761 blue line withldquo998779rdquo represents for the problem with objective 1198762

63 Convergence of Optimization Algorithms In this sectionwe address the convergence rate of our optimization algo-rithms by investigating how the parameters of network size119873 and the maximal possible transmit power 119875max can impactthe convergence rate

It is observed from Figure 4 that the algorithm for theproblem with objective 1198762 quickly converges to the optimalsolution while the algorithm for the problem with objective1198761 converges to the optimal solution at a low rate Indeed theformer can reach the optimum after 34times104 iterations while



0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()

20 30 40 5010Size of network(b)

Figure 5 The figure illustrates the distribution of users in each group with network size 119873 (a) Distribution of users with objective 1198761(b) distribution of users with objective 1198762 Green line with ldquoIrdquo represents the percentage of users in 1198661 red line with ldquordquo represents thepercentage of users in 1198662 blue line with ldquo998779rdquo represents the percentage of users in 1198663

the latter convergence appears after 23times104 iterations undera network of mobile hospital at a scale of 50

64Distribution ofUsers amongThreeGroups In this sectionwe present the distribution of users among three groups at theoptimum and investigate the impact of optimization policies(the objectives of optimization problems) on the distribution

It is observed from Figure 5 that the distribution of usersin three groups is quite different for the objective of 1198761 and1198762 In the optimal solution to the problem with objective1198761the percentage of users in1198661 increases dramatically while thepercentage of users in1198663 decreases dramatically with the riseof119873The percentage of users in1198662 is close to 0 In the optimalsolution to the problem with objective 1198762 the percentage ofusers in 1198662 increases dramatically while the percentage ofusers in 1198663 decreases dramatically with the rise of 119873 Thepercentage of users in 1198661 is close to 0

In the results for both objectives 1198761 and 1198762 the per-centage of users in 1198663 decreases with the rise of 119873 In otherwords the base station starts to reduce the transmit powerof most users This is because a large number of users canlead to a great amount of interference to any user and thusthe base station must decrease the transmit power to reducethe interference within the whole network

7 Conclusion

We consider the setting of optimizing QoS in a network ofmobile hospital in which a user receives both signal frombase station and the noise from the other usersWith the QoS

metric (ie SNR) we study the optimization of QoS withinthe whole network and propose a novel algorithm to allocatethe transmit power for themaximumof SNR Some of the keyinferences drawn are

(i) the problems with two proposed objectives can yieldto two policies of optimizing QoS a greedy policy isoptimal to maximize the sum of SNR of each userwhereas a fair policy is optimal to maximize theproduct of SNR of each user

(ii) proposed SNR optimization algorithm can achievethe optimum within a running time which linearlyincreases with the size of network119873

(iii) the analysis on SNR optimization shows the partitionof users into three groups with certain features andhow the parameters of network impact the distribu-tion of users among these groups

Wewould also like to extend our results to a setting whichcontains multiple priorities of users and in such a settinga few users in the higher priority may have more rigorousQoS requirements than low-priority users Also we wouldlike to investigate how to design QoS-optimization policiesunder a setting when users are noncooperative and each usermay reject the allocation of transmit power with a certainprobability

Conflict of Interests

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


Acknowledgments

This work was partially supported by the National Natu-ral Science Foundation of China (no 61370202) partiallysupported by a grant from the National High TechnologyResearch and Development Program of China (863 Programno 2012AA02A614) and partially supported by the Fun-damental Research Funds for the Central Universities (noZYGX2015J071)

References

[1] P Phunchongharn D Niyato E Hossain and S CamorlingaldquoAn EMI-aware prioritized wireless access scheme for e-Healthapplications in hospital environmentsrdquo IEEE Transactions onInformation Technology in Biomedicine vol 14 no 5 pp 1247ndash1258 2010

[2] P Phunchongharn E Hossain and S Camorlinga ldquoElec-tromagnetic interference-aware transmission scheduling andpower control for dynamic wireless access in hospital envi-ronmentsrdquo IEEE Transactions on Information Technology inBiomedicine vol 15 no 6 pp 890ndash899 2011

[3] Y L Che R Zhang Y Gong and L Duan ldquoOn spatial capacityof wireless ad hoc networks with threshold based schedulingrdquoIEEE Transactions on Wireless Communications vol 13 no 12pp 6915ndash6927 2014

[4] Y L Che R Zhang Y Gong and L Duan ldquoRobust optimiza-tion of cognitive radio networks powered by energy harvest-ingrdquo in Proceedings of the 34th IEEE International Conferenceon Computer Communications (INFOCOM rsquo15) Hong KongApril-May 2015

[5] Q Shen X Liang X Shen X Lin and H Y Luo ldquoExploitinggeo-distributed clouds for a E-health monitoring system withminimum service delay and privacy preservationrdquo IEEE Journalof Biomedical and Health Informatics vol 18 no 2 pp 430ndash4392014

[6] M R Javan and A R Sharafat ldquoEfficient and distributed SINR-Based joint resource allocation and base station assignment inwireless CDMA networksrdquo IEEE Transactions on Communica-tions vol 59 no 12 pp 3388ndash3399 2011

[7] K-S Tan and I Hinberg ldquoRadiofrequency susceptibility testson medical equipmentrdquo in Proceedings of the 16th AnnualInternational Conference of the IEEE Engineering in Medicineand Biology Society Engineering Advances New Opportunitiesfor Biomedical Engineers vol 2 pp 998ndash999 Baltimore MdUSA November 1994

[8] ldquoElectromagnetic compatibility of medical devices with mobilecommunicationsrdquo inMedical Devices Bulletin DB9702 MedicalDevices Agency London UK 1997

[9] A J Trigano A AzoulayM Rochdi andA Campillo ldquoElectro-magnetic interference of external pacemakers by walkie-talkiesand digital cellular phones experimental studyrdquo Pacing andClinical Electrophysiology vol 22 no 4 pp 588ndash593 1999

[10] G Calcagnini P Bartolini M Floris et al ldquoElectromagneticinterference to infusion pumps from GSM mobile phonesrdquoin Proceedings of the 26th Annual International Conference ofthe IEEE Engineering in Medicine and Biology Society (IEMBSrsquo04) vol 2 pp 3515ndash3518 IEEE San Francisco Calif USASeptember 2004

[11] Y Chu and A Ganz ldquoA mobile teletrauma system using 3Gnetworksrdquo IEEE Transactions on Information Technology inBiomedicine vol 8 no 4 pp 456ndash462 2004

[12] E A V Navarro J R Mas J F Navajas and C P AlcegaldquoPerformance of a 3G-based mobile telemedicine systemrdquo inProceedings of the 3rd IEEE Consumer Communications andNetworking Conference (CCNC rsquo06) vol 2 pp 1023ndash1027 IEEELas Vegas Nev USA January 2006

[13] J Shepherd 2009 httpwwwtheguardiancomsociety2009jan06mobile-phones-hospitals

[14] C-K Tang K-H Chan L-C Fung and S-W Leung ldquoElectro-magnetic interference immunity testing of medical equipmentto second- and third-generationmobile phonesrdquo IEEE Transac-tions on Electromagnetic Compatibility vol 51 no 3 pp 659ndash664 2009

[15] M Ardavan K Schmitt and C W Trueman ldquoA preliminaryassessment of EMI control policies in hospitalsrdquo in Proceedingsof the 14th International Symposium on Antenna Technology andApplied Electromagnetics and the American ElectromagneticsConference (ANTEMAMEREM rsquo10) pp 1ndash6 Ottawa CanadaJuly 2010

[16] S Krishnamoorthy J H Reed C R Anderson P N Robertand S Srikanteswara ldquoCharacterization of the 24GHz ISMband electromagnetic interference in a hospital environmentrdquoin Proceedings of the 25th Annual International Conference ofthe IEEE Engineering in Medicine and Biology Society (EMBCrsquo03) vol 4 pp 3245ndash3248 IEEE Buenos Aires ArgentinaSeptember 2003

[17] D Witters and S Seidman ldquoEMC and wireless healthcarerdquo inProceedings of the Asia-Pacific Symposium on ElectromagneticCompatibility Beijing China April 2010

[18] S GMyerson ldquoMobile phones in hospitals are not as hazardousas believed and should be allowed at least in non-clinical areasrdquoBritish Medical Journal vol 326 no 7387 pp 460ndash461 2003

[19] F Fiori ldquoIntegrated circuit susceptibility to conducted RFInterferencerdquo Compliance Engineering vol 17 pp 40ndash49 2014

[20] D Lin X Wu F Labeau and A Vasilakos ldquoInternet of vehiclesfor E-health applications in view of EMI on medical sensorsrdquoJournal of Sensors vol 2015 Article ID 315948 10 pages 2015

[21] W D Kimmel and D D Gerke Ten Common EMI Problems inMedical Electronics 2005 httpwwwmedicalelectronicsdesigncomarticleten-common-emi-problems-medical-electronics

[22] G Acampora D J Cook P Rashidi and A V Vasilakos ldquoAsurvey on ambient intelligence in healthcarerdquo Proceedings of theIEEE vol 101 no 12 pp 2470ndash2494 2013

[23] D He C Chen S Chan J Bu and A V Vasilakos ldquoReTrustattack-resistant and lightweight trust management for medicalsensor networksrdquo IEEE Transactions on Information Technologyin Biomedicine vol 16 no 4 pp 623ndash632 2012

[24] N Xiong A V Vasilakos L T Yang et al ldquoComparativeanalysis of quality of service and memory usage for adaptivefailure detectors in healthcare systemsrdquo IEEE Journal on SelectedAreas in Communications vol 27 no 4 pp 495ndash509 2009

[25] ITU-R Recommendation M1225 ldquoGuidelines for evaluation ofradio transmission technologies for IMT-2000rdquo Question ITU-R 398 1997

[26] J Leskovec K J LangADasgupta andMWMahoney ldquoCom-munity structure in large networks natural cluster sizes and theabsence of large well-defined clustersrdquo Internet Mathematicsvol 6 no 1 pp 29ndash123 2009

Submit your manuscripts athttpwwwhindawicom

Computer Games Technology

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Distributed Sensor Networks


Advances in

FuzzySystems

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014


ReconfigurableComputing

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014


Applied Computational Intelligence and Soft Computing

thinspAdvancesthinspinthinsp

Artificial Intelligence

HindawithinspPublishingthinspCorporationhttpwwwhindawicom Volumethinsp2014

Advances inSoftware EngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Electrical and Computer Engineering

Journal of

Journal of

Computer Networks and Communications


Hindawi Publishing Corporation

httpwwwhindawicom Volume 2014

Advances in

Multimedia


Biomedical Imaging


ArtificialNeural Systems

Advances in


RoboticsJournal of



Computational Intelligence and Neuroscience

Industrial EngineeringJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Human-ComputerInteraction

Advances in

Computer EngineeringAdvances in



a high level of power may lead to the malfunction of medicalsensors Such an improper power allocation by the above-mentioned algorithms may influence the function of EMI-sensitive medical sensors so these algorithms cannot beemployed under the scenario of mobile hospital Also theabove-mentioned algorithms are designed to optimize theindividual objective of each wireless user instead of opti-mizing a network-level objective (eg the quality of networkservice) The importance of scheduling wireless transmissionunder a mobile hospital scenario and the lack of efficientalgorithms for optimizing network-level objective motivate usto study how wireless users can control their transmit powerto achieve certain goals such as maximizing the quality of net-work service while ensuring the acceptable level of EMI onmed-ical sensors over Internet of vehicles for E-health applications

12 Main Contributions In this paper we present the issueof scheduling wireless transmission under the environmentof mobile hospital The objective of this paper is to optimizecertain goals (eg the quality of service) at the network levelinstead of at the user level In this paper we address theproblem of optimizing the quality of network service in amobile hospital environment and propose the algorithm ofpower control to achieve the optimal network service To thebest of our knowledge this is the first work which addressesthe algorithm of power control to achieve a network-levelgoal under a wireless network for E-health applications Theprimary contributions of this paper are composed of the fol-lowing issues (i) establishing a problem of optimizing powercontrol to achieve the globally optimal quality of service atthe network level in view of EMI on medical sensors (ii)analyzing the transmit power and quality of service of eachuser at the optimal solution to the proposed problem

2 Related Work of EMI on Medical Sensors

In hospital scenarios the research on EMI begins with thestudy of immunity of medical equipment to mobile phonesTan and Hinberg in [7] address that a few pieces of medicalequipment are sensitively influenced by the EMI frommobilephones and the typical pieces of equipment include infusionpumps ventilators and ECGmonitors Also an EMI suscep-tibility test was conducted by the Medicines and HealthcareProducts Regulatory Agency (MHRA) of UK [8] This testfocuses on investigating the EMI of mobile phones within apersonal communication network and its results show thatthe EMI-sensitive medical equipment include respiratorsdefibrillators and external pacemakers Trigano et al in [9]and Calcagnini et al in [10] present how the EMI fromGSM mobile phones influences pacemakers and infusionpumps respectively The results show that the EMI of mobilephones can lead to the malfunction of both pacemakers andinfusion pumps With the wide use of 3rd-generation (3G)telecommunication systems all over the world the authors in[11 12] study how EMI impacts medical pieces of equipmentwithin the 3G bands Based on the aforementioned researchthe International Electrotechnical Committee (IEC) in 2007publishes the EN60601-1-2 standard and this standard rec-ommends the level of EMI immunity as 10Vm and 3Vm

for non-life-supporting equipment (eg defibrillators) andlife-supporting equipment (eg blood pressuremonitors andinfusion pumps) respectively In view of the advances ofelectromagnetic compatibility (EMC) technologies Singa-pore and the UK governments relax the EMI restrictionwhich is recommended by EN60601-1-2 standard andmobilephones are allowed to be uses in certain areas of hospitals[13] In consideration of the most advanced EMC of medicalequipment Tang et al in [14] address an EMI test whichinvestigates the EMI fromGSM900 PCS1800 and 3Gmobilecommunication systems This test shows that EMI-sensitiveequipment includes ECG monitors audio evoked potentialsystems radiographic systems and ultrasonic fetal heartdetectors [14] From the study results of previous literaturewe can reach a conclusion syringe pumps ECG monitorsfetal monitors respirators anesthesia machines externalpacemakers infusion pumps and defibrillators are sensitiveto the EMI of mobile phones [15]

The other studies focus on the EMI from devices whichhave access to a wireless local area network (WLAN) anda typical WLAN usually works around the frequency of24GHz which is different from the working band of mobilephones The amount of EMI on medical equipment is par-tially determined by frequency bands and thus the researchon EMI under the scenario of healthcare monitoring withina WLAN appears Krishnamoorthy et al in [16] carry out ameasurement of EMI caused by doctor devices on pieces ofmedical equipment located in three hospitals and the doctordevices work within a 24GHz-band WLAN The measure-ments show that the highest level of EMI is 0552Vmwhich is in the acceptable EMI range recommended by theEN60601-1-2 standard However the test in [16] did not takeinto account the quality of service (QoS) of data whichare transmitted between patient devices and doctor devicesThe strategies of restricting the use of mobile phones suchas switching off mobile phones are not applicable for thescenario of a wireless healthcare monitoring system [17] Inwireless healthcare monitoring systems doctors and patientsneed to employwireless devices to transmit data and establishcommunication and the restriction on the level of transmitpower may cause a lower level of QoS of data transmissionwhich would lead to the loss ofmedical dataThus in wirelesshealthcare monitoring systems the restriction of transmitpower and QoS requirements are usually against each otherAdditionally the simultaneous transmission of data amongmultiple patient devices and doctor devices would cause ahigh level of EMI to medical equipment [1] Phunchongharnet al in [1] investigate the EMI in hospital environments byconsidering the QoS of patient devices and doctor devicesand they conclude that the level of EMI on most of themedical equipment is unacceptable when the transmit powerof a device is above 10mW within a WLAN

All the above-mentioned research does not consider thevehicular scenarios for healthcare applications which areinteresting to this paper and thus the medical sensors inthe test may not be vehicle-mounted and wearable medicalsensors In Section 31 we address a detailed experimentwhich includes the test of EMI impact on types of vehicle-mounted and wearable medical sensors










Mobile station

Holter



Holter


EMI impact

Hospital


over phone)







sum119895isin119878

1205791radic119875119895119863119895 (119909)

le 119864NL (119909) for 119909 isin 1198781 (1)

sum119895isin119878

1205792radic119875119895119863119895 (119910)

le 119864L (119910) for 119910 isin 1198782 (2)







SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868 (3)



1198761 = sum119894

SNR119894 (4)

1198762 = prod119894

SNR119894 (5)




maxP

sum119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(6)

maxP

prod119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(7)



maxP

sum119894

119909119894sum119895 =119894 119909119895 + 119868


119895 =119894

119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(8)


maxP

prod119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894


119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(9)



119871 = 119876 minussum119894

120579119894 (119909119894ℎ119894minus 119875max)

minussum119894


119909119895)

minus 120583(sum119894

119909119894 minus 119875119878)

= 119876 minussum119894


119895 =119894

120582119895)119909119894

+ Const

(10)



120597119871120597119909119894

= 0 997888rarr


119895 =119894

120582119895 =120597119876120597119909119894

(11)








Proof Given

1205971198761120597119909119894

= 1119868 + sum119873119896=1119896 =119894 119909119896

minus119873

sum119898=1119898 =119896

119909119898(119868 + sum119873119896=1119896 =119898 119909119896)

2 (12)

we have

1205971198761120597119909119894

minus 1205971198761120597119909119895

= (119868 +119873

sum119896=1119896 =119894

119909119896)

sdot

1

(119868 + sum119873119896=1119896 =119894 119909119896)2minus 1

(119868 + sum119873119896=1119896 =119895 119909119896)2

(13)








120598

0(12059711987611205971205980)1198891205980 gt 0 since 12059711987611205971205980 = (1205971198761120597119909119894 minus

1205971198761120597119909119895)|x=x1205980




lowast119894 119909

lowast119895 119909






lowast119894 119909








lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198761120597119909


119909lowast119894 lt 119909lowast119895








Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)




1

120582119896 So 119909119894 = 119909119895 by Lemma 10



lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909


119909lowast119894 lt 119909lowast119895









end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in












Mobile station

Holter



Holter


EMI impact

Hospital


over phone)







sum119895isin119878

1205791radic119875119895119863119895 (119909)

le 119864NL (119909) for 119909 isin 1198781 (1)

sum119895isin119878

1205792radic119875119895119863119895 (119910)

le 119864L (119910) for 119910 isin 1198782 (2)







SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868 (3)



1198761 = sum119894

SNR119894 (4)

1198762 = prod119894

SNR119894 (5)




maxP

sum119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(6)

maxP

prod119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(7)



maxP

sum119894

119909119894sum119895 =119894 119909119895 + 119868


119895 =119894

119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(8)


maxP

prod119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894


119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(9)



119871 = 119876 minussum119894

120579119894 (119909119894ℎ119894minus 119875max)

minussum119894


119909119895)

minus 120583(sum119894

119909119894 minus 119875119878)

= 119876 minussum119894


119895 =119894

120582119895)119909119894

+ Const

(10)



120597119871120597119909119894

= 0 997888rarr


119895 =119894

120582119895 =120597119876120597119909119894

(11)








Proof Given

1205971198761120597119909119894

= 1119868 + sum119873119896=1119896 =119894 119909119896

minus119873

sum119898=1119898 =119896

119909119898(119868 + sum119873119896=1119896 =119898 119909119896)

2 (12)

we have

1205971198761120597119909119894

minus 1205971198761120597119909119895

= (119868 +119873

sum119896=1119896 =119894

119909119896)

sdot

1

(119868 + sum119873119896=1119896 =119894 119909119896)2minus 1

(119868 + sum119873119896=1119896 =119895 119909119896)2

(13)








120598

0(12059711987611205971205980)1198891205980 gt 0 since 12059711987611205971205980 = (1205971198761120597119909119894 minus

1205971198761120597119909119895)|x=x1205980




lowast119894 119909

lowast119895 119909






lowast119894 119909








lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198761120597119909


119909lowast119894 lt 119909lowast119895








Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)




1

120582119896 So 119909119894 = 119909119895 by Lemma 10



lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909


119909lowast119894 lt 119909lowast119895









end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in









SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868 (3)



1198761 = sum119894

SNR119894 (4)

1198762 = prod119894

SNR119894 (5)




maxP

sum119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(6)

maxP

prod119894

119875119894ℎ119894sum119895 =119894 119875119895ℎ119895 + 119868

st 0 le 119875119894 le 119875max

SNR119894 =119875119894ℎ119894

sum119895 =119894 119875119895ℎ119895 + 119868ge SNRmin

sum119894

119875119894ℎ119894 le 119875119878

(7)



maxP

sum119894

119909119894sum119895 =119894 119909119895 + 119868


119895 =119894

119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(8)


maxP

prod119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894


119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(9)



119871 = 119876 minussum119894

120579119894 (119909119894ℎ119894minus 119875max)

minussum119894


119909119895)

minus 120583(sum119894

119909119894 minus 119875119878)

= 119876 minussum119894


119895 =119894

120582119895)119909119894

+ Const

(10)



120597119871120597119909119894

= 0 997888rarr


119895 =119894

120582119895 =120597119876120597119909119894

(11)








Proof Given

1205971198761120597119909119894

= 1119868 + sum119873119896=1119896 =119894 119909119896

minus119873

sum119898=1119898 =119896

119909119898(119868 + sum119873119896=1119896 =119898 119909119896)

2 (12)

we have

1205971198761120597119909119894

minus 1205971198761120597119909119895

= (119868 +119873

sum119896=1119896 =119894

119909119896)

sdot

1

(119868 + sum119873119896=1119896 =119894 119909119896)2minus 1

(119868 + sum119873119896=1119896 =119895 119909119896)2

(13)








120598

0(12059711987611205971205980)1198891205980 gt 0 since 12059711987611205971205980 = (1205971198761120597119909119894 minus

1205971198761120597119909119895)|x=x1205980




lowast119894 119909

lowast119895 119909






lowast119894 119909








lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198761120597119909


119909lowast119894 lt 119909lowast119895








Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)




1

120582119896 So 119909119894 = 119909119895 by Lemma 10



lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909


119909lowast119894 lt 119909lowast119895









end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




maxP

prod119894

119909119894sum119895 =119894 119909119895 + 119868

st 0 le 119909119894 le 119875maxℎ119894


119909119895 le 0

sum119894

119909119894 minus 119875119878 le 0

(9)



119871 = 119876 minussum119894

120579119894 (119909119894ℎ119894minus 119875max)

minussum119894


119909119895)

minus 120583(sum119894

119909119894 minus 119875119878)

= 119876 minussum119894


119895 =119894

120582119895)119909119894

+ Const

(10)



120597119871120597119909119894

= 0 997888rarr


119895 =119894

120582119895 =120597119876120597119909119894

(11)








Proof Given

1205971198761120597119909119894

= 1119868 + sum119873119896=1119896 =119894 119909119896

minus119873

sum119898=1119898 =119896

119909119898(119868 + sum119873119896=1119896 =119898 119909119896)

2 (12)

we have

1205971198761120597119909119894

minus 1205971198761120597119909119895

= (119868 +119873

sum119896=1119896 =119894

119909119896)

sdot

1

(119868 + sum119873119896=1119896 =119894 119909119896)2minus 1

(119868 + sum119873119896=1119896 =119895 119909119896)2

(13)








120598

0(12059711987611205971205980)1198891205980 gt 0 since 12059711987611205971205980 = (1205971198761120597119909119894 minus

1205971198761120597119909119895)|x=x1205980




lowast119894 119909

lowast119895 119909






lowast119894 119909








lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198761120597119909


119909lowast119894 lt 119909lowast119895








Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)




1

120582119896 So 119909119894 = 119909119895 by Lemma 10



lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909


119909lowast119894 lt 119909lowast119895









end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198761120597119909


119909lowast119894 lt 119909lowast119895








Proof Given

1205971198762120597119909119894

= 1198762 1119909119894minus119873

sum119898=1119898 =119894

1(119868 + sum119873119896=1119896 =119898 119909119896)

(14)

we have

1205971198762120597119909119894

minus 1205971198762120597119909119895

= 1119909119894 (119868 + 119909119894)

minus 1119909119895 (119868 + 119909119895)

(15)




1

120582119896 So 119909119894 = 119909119895 by Lemma 10



lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the




lowast119895 119909

lowast119873 is the


lowast119894 minus 1205971198762120597119909


119909lowast119894 lt 119909lowast119895









end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






end if119906 = 119906 minus Δ119875

end while119879 = 119879 + 1











end if119906 = 119906 + Δ119875

end while119879 = 119879 + 1









+ 80(16)



0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




0

01

02

03

04

05

06

07

08

09

Aver

age l

evel

of S

NR


(a)


0

005

01

015

02

025

Stan

dard

dev

iatio

n of

SN

R(b)










Averagepower (dB)

Dopplerspectrum





0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




0

005

01

015

02

025

Aver

age l

evel

of S

NR


(a)

0

01

02

03

04

05

06

Aver

age l

evel

of S

NR

04 06 08021I

(b)


times104

1

15

2

25

3

Num

ber o

f ite

ratio

ns


(a)

times104


2

22

24

26

28

3

32

34

Num

ber o

f ite

ratio

ns

(b)






0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in





0

20

40

60

80

100

Use

rs in

a gr

oup

()

(a)

0

20

40

60

80

100

Use

rs in

a gr

oup

()







7 Conclusion










Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




Acknowledgments


References


































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in










Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in



Research Article Optimal Network QoS over the...

Documents

Transcript of Research Article Optimal Network QoS over the...