Properties of Indoor Received Signal Strength for WLAN Location Fingerprinting

Size: px
Start display at page:

Download "Properties of Indoor Received Signal Strength for WLAN Location Fingerprinting"

Transcription

1 Propertes of Indoor Receved Sgnal Strength for WLAN Locaton Fngerprntng Kamol Kaemarungs and Prashant Krshnamurthy Telecommuncatons Program, School of Informaton Scences, Unversty of Pttsburgh E-mal: Abstract Indoor postonng systems that make use of receved sgnal strength based locaton fngerprnts and exstng wreless local area network nfrastructure have recently been the focus for supportng locatonbased servces n ndoor and campus areas. A knowledge and understandng of the propertes of the locaton fngerprnt can assst n mprovng desgn of algorthms and deployment of poston locaton systems. However, most exstng research work gnores the rado sgnal propertes. Ths paper nvestgates the propertes of the receved sgnal strength reported by IEEE 82.b wreless network nterface cards. Analyses of the data are performed to understand the underlyng features of locaton fngerprnts. The performance of an ndoor postonng system n terms of ts precson s compared usng measured data and a Gaussan model to see how closely a Gaussan model may ft the measured data. Keywords: ndoor, measurement, modelng, postonng system, wreless LANs. Introducton Indoor postonng systems that use locaton fngerprnts and exstng wreless local area network (WLAN) nfrastructure have been demonstrated for ndoor areas [] where the global postonng system (GPS) does not work well [2]. The fngerprntng technque s smple to deploy compared to technques usng angle of arrval (AOA) and tme dfference of arrval (TDOA). Instead of dependng on accurate estmates of angle or dstance to determne the locaton, locaton fngerprntng assocates locatondependent characterstcs such as the receved sgnal strength (RSS) wth a locaton and uses these characterstcs to nfer the locaton. In ths case, there s no need for specalzed hardware at the moble staton (MS) besdes the wreless network nterface card (NIC) and the exstng WLAN nfrastructure can be reused easly. Before a postonng system can estmate the locaton, a locaton fngerprnt database or a rado map [3] must be constructed. Each entry n the database s a mappng between a poston and a locaton fngerprnt. The locaton fngerprnt can be an average value as n the RADAR system [] or probablstc [3]. In the average approach that we consder n ths paper, the locaton fngerprnt s a vector R of the average RSS values from multple access ponts (APs) at a partcular locaton L. A typcal vector R = (r, r 2,, r N ) conssts of N RSS values from N APs. The rado map contans all such RSS vectors for a grd of locatons n the ndoor area. For postonng, a MS obtans a sample RSS vector P = (, 2,, N ). The Eucldean sgnal dstance between the P and R for each R n the database s computed. The locaton s then estmated to be that L for whch the Eucldean dstance s the smallest. Note that the vector P s random. An error s made when the smallest Eucldean dstance occurs for a locaton L that s not the one at whch the sample P was collected. Errors occur because the measured RSS vector s a sample of a random vector whle only the average RSS vector s stored n the rado map. Understandng the statstcal propertes of the locaton fngerprnt (RSS vector) s mportant for the desgn of postonng systems for several reasons. It can provde nsghts nto how many access ponts are needed to unquely dentfy a locaton [3] wth a gven accuracy and precson, whether preprocessng of the RSS measurements can mprove the accuracy and so on. Exstng lterature on ndoor postonng systems focuses manly on the accuracy performance and mprovement of the locaton estmaton algorthm and gnores the study of the RSS random vector. Knowledge of the RSS propertes can n fact enable the development of better algorthms to classfy a measured RSS vector P as belongng to a partcular

2 z z locaton L. Although a varety of statstcal rado propagaton models exst, they were developed wth sgnal coverage, communcatons capablty and data rate n mnd. Moreover, the relatonshp between RSS values from multple APs s not understood very well. The dstrbuton of RSS values, ther standard devaton, ther temporal varaton, and the (n)dependence of RSSs from multple access ponts (APs) are mportant for understandng and modelng the performance of fngerprnt based ndoor postonng systems. For nstance, the dstrbuton of the RSS s sad to be normally dstrbuted n dbm accordng to the study n [5]. However, our prelmnary study and the study n [6] showed otherwse. Ths artcle presents data analyses of the RSS n an ndoor envronment wth postonng n mnd. Secton 2 descrbes the measurement setup. Secton 3 explores the effect of user s body, the effect of user s orentaton, the statonarty and tme-dependence, and the dstrbuton of the RSS. The ndependence of the RSS from multple APs s also analyzed n Secton 3. An approxmate model for the RSS s appled to the study of the precson performance of postonng systems n Secton 4 to evaluate the sutablty of a Gaussan approxmaton. We conclude the paper n Secton Measurement setup A standard laptop computer equpped wth an Ornoco WLAN card and clent manager software was used to collect samples of RSS from APs nsde the School of Informaton Scences (IS) buldng at the Unversty of Pttsburgh. The WLAN card s plugged nto the PCMCIA slot on the rght sde of the laptop. The buldng has 8 floors and APs nstalled opportunstcally. The dmenson of each floor s approxmately 76 ft 2 ft (23 m 37 m). All APs are from Lucent s WAVELAN and are equpped wth Ornoco WLAN cards. The rado frequency channels of IEEE 82.b are n the 2.4 GHz band whch s shared by other equpment n the ndustral, scentfc, and medcal (ISM) band such as Bluetooth. The number of non-overlappng channels for 82.b s three [7]. We observe that the RSS value reported by the WLAN card s an average value over a samplng perod and n ntegral steps of dbm. The receved sgnal senstvty of the WLAN card also lmts the range of the RSS to be between -93 dbm and dbm [8]. Nevertheless, the hghest typcal value of the RSS s approxmately -3 dbm at one meter from any AP. The clent manager software s a ste survey tool whch provdes the lnk qualty and AP montorng capabltes. 2.. Expermental desgn The measurement n each of the studes n Secton 3.2 s done by samplng the RSS data every one second. The vector of RSS data at each locaton forms the locaton fngerprnt wth at most three RSS elements n the vector. Four locatons of measurement are chosen on the fourth floor of the IS buldng as shown n Fgure denoted as L, L2, L3, and L4. The user s orentaton corresponds to the arrows at each locaton n ths fgure. The locatons of APs on ths floor are labeled as SIS4, SIS4, and SIS48. The rado channels used for each AP are channel, 6, and, respectvely. There s one more AP on the ffth floor labeled as SIS5 wth channel number 6, but there s no AP on the thrd floor. The MS has a drect lne-ofsght to one of the APs only at L2. The data were collected four tmes wth a perod of approxmately one hour each for every locaton and at dfferent hours of the day. The total number of RSS samples would be 4 locatons 4 hours 3 APs = 48. However, n our experments only 46 RSS data samples were collected because at L we can receve sgnals from only two APs at best SIS GIS Lab North UP DN 4a L 496 SIS 4 4-Wreless Lab L2 L3 z UP SIS 5 DN SIS L4 z 4 Fgure. The fourth floor of the IS buldng wth the locatons of APs and the orentaton of measurements 3. Propertes of the receved sgnal strength Indoor rado propagaton s dffcult to predct because of the dense multpath envronment and propagaton effects such as reflecton, dffracton, and scatterng [9]. Multpath fadng causes the receved sgnal to fluctuate around a mean value at partcular locaton. The receved sgnal s usually modeled by the combned effects of large-scale fadng and small-scale fadng []. The large-scale-fadng component (of nterest here) descrbes the sgnal attenuaton as the 42

3 sgnal travels over a dstance and s absorbed by materal such as walls and floors along the way to the recever. Ths component predcts the mean of the RSS and usually has a log-normal dstrbuton []. Small-scale fadng explans the dramatc fluctuaton of the sgnal due to multpath fadng. If there s no lneof-sght (NLOS) component, the small-scale fadng s often modeled wth a Raylegh dstrbuton. If there s a lne-of-sght (LOS) component, the small-scale fadng s modeled wth a Rcan dstrbuton. However, these models are focused on understandng the mpact of rado propagaton on recever desgn and sgnal coverage rather than from the perspectve of ndoor postonng systems. The nvestgaton of RSS data n ths secton s dvded nto four parts. All measurements were done at fxed locatons. Frst, we consder the effect of the user s presence on a sngle RSS set (ths s the set of RSS samples from one AP at a fxed locaton obtaned over tme). Second, we nvestgate the statstcal propertes of a sngle RSS set (the dstrbuton, the statonarty, the tme-of-day dependency, etc.). Thrd, we study the propertes of multple RSS sets (bascally RSS values from multple APs). We evaluate whether each RSS set s ndependent from the others and whether they all exhbt the same statstcal propertes. We also observed that n some locatons the sgnal from certan APs s not present all the tme. Fnally, we compare the dfferences between the RSS fngerprnts (these are vectors wth RSS sets as ther components) of two locatons. 3.. Effects of user s presence on RSS In ndoor postonng systems based on WLANs, the user typcally carres the moble staton equpped wth a wreless NIC. The effect of the user s presence close to the antenna plays a sgnfcant role n the mean value and the spread of the average RSS values. An observaton was made n [] that the user s orentaton caused a varaton n RSS level up to 5 dbm Effect of user s body. To study the effect of the user s body, we performed measurement of the sgnal from SIS4 at locaton L nsde the room IS 4a n Fgure. The dstance between the transmtter (AP) and the recever (MS) s approxmately 7 m and the MS does not have a clear lne-of-sght to the AP. The data were recorded for two hours. Durng frst hour, the user was present, whle no user was present n the second hour. The results were analyzed by plottng hstograms of the RSS for both hours. The results are shown n Fgure 2. Frequency Frequency (a) Dstrbuton of RSS when user s present Average receved sgnal strength n dbm (b) Dstrbuton of RSS when user s not present Average receved sgnal strength n dbm Fgure 2. Comparson of hstograms of RSS Fgure 2-a and 2-b depct the dfference between these two dstrbutons. The user s body nfluences the RSS dstrbuton by spreadng the range of RSS values by a sgnfcant amount. The standard devaton s ncreased from approxmately.68 dbm to 3. dbm when the user s present. The mean changes from dbm to -7.6 dbm wth the user s body present. Clearly, t s essental to collect data for the rado map based on the applcaton. When the postonng system s supposed to cater to real users, t s essental to have the user present whle collectng the RSS values for the fngerprnt and to take nto account the effect of human s body. For applcatons that make use of sensors wthout a human presence the data should reflect that envronment Effect of user s orentaton. Because the resonance frequency of water s at 2.4 GHz and the human s body conssts of 7% water, the RSS s absorbed when the user obstructs the sgnal path and causes an extra attenuaton [6]. To study the effect of user s orentaton, we performed another measurement at the locaton L2 nsde the room IS 4 n Fgure. In ths case, there s a lne-of-sght between the transmtter (SIS4) and recever and the dstance between them s approxmately 2 ft (6 m). Sgnals from SIS4 and SIS5 were also present at ths locaton wth non lne-of-sght dstances of 36 ft ( m) and 22 ft (7 m), respectvely. The measurement was done wth four orentatons (facng North, West, South, and East of the buldng) for a perod of 5 mnutes each. The results of the statstcs of the RSS values from the three transmtters are shown n Tables, 2, and 3. The orentatons that user body blocks the drect path between the AP and MS are marked wth astersks.

4 Table. LOS RSS (dbm) from SIS4 wth dfferent orentaton Statstcs North West South* East Sample Mean Standard Devaton Skewness For the LOS case (Transmtter SIS4) n Table, when the user was facng south and the AP was behnd the user, the sample mean of the RSS was lower at dbm compared to the hghest RSS of dbm when the user faced west. The results show that the RSS can be attenuated by 9.32 db n our case due to the obstructon from the body. Ths suggests that the user s orentaton s crucal and should be ncluded n computng the user locaton nformaton as ponted out n []. The attenuaton by the body of the user can even completely block the RSS from a NLOS AP as shown n Table 2 when there was no RSS nformaton at all durng the perod that the person s back was turned towards the transmtter SIS4. Ths means that the locaton fngerprnt at the same locaton may lack one RSS value n the vector f the user s orentaton s dfferent. The sgnal from SIS5 s also attenuated by 5.8 db between the hghest and the lowest RSS levels n Table 3. Table2. NLOS RSS (dbm) from SIS4 wth dfferent orentaton Statstcs North West* South East Sample Mean N/A Standard Devaton.84 N/A Skewness.5 N/A.9 -. Table3. NLOS RSS (dbm) from SIS5 wth dfferent orentaton Statstcs North West* South East Sample Mean Standard Devaton Skewness In what follows, we restrct the study and focus on the RSS propertes when the user s present and facng one drecton arbtrarly Statstcal propertes of the RSS Tradtonally, the average RSS s beleved to be log-normally dstrbuted accordng to popular largescale fadng models []. The mean value s generally predctable and beleved to follow one of several standardzed path loss models dscussed n [9]. However, there are some conflctng conclusons regardng the RSS dstrbuton measured at the software level by the wreless NIC for ndoor rado propagaton n [5] and [6]. Moreover, the standard devaton and the statonarty of the RSS are not understood very well Dstrbuton of receved sgnal strength. The results n [5] are based on a fve second samplng perod over long duratons of fve hours, 2 hours, and one month. Here, they conclude that the RSS s lognormally dstrbuted (normal or Gaussan n db) due to the smlarty of the medan, the mean, and the mode. However, they dd not ndcate whether the user was present all the tme durng the measurements. Thus, we suspect that the dstrbuton of the RSS n db that could be observed n realty may not be normally dstrbuted as descrbed n [5]. A recent study of a 45- second measurement perod wth the user s presence n [6] ponted out that the RSS dstrbuton was non- Gaussan and asymmetrc. Moreover, the hstograms n [6] depcted that there could be multple modes wth one domnant mode n the dstrbuton. The means and the modes were often dfferent n ther results. The results of our experment showed a smlar trend as n [6]. The hstogram n Fgure 2-a has two modes wth one domnant mode when the user s present. A vsual test of Fgure 2-a confrms that the RSS does not come from a normal dstrbuton and a norm-plot test s also nonlnear. Out of the 46 RSS elements that we collected from four locatons (L to L4), most of the hstograms show that the RSS does not ft the normal dstrbuton. Only a few hstograms showed a good normal approxmaton. We observed that the RSS dstrbuton tended to be left-skewed n our measurement results. The hstograms that are strongly left-skewed are usually the ones wth the strongest RSS out of the three APs at a partcular locaton. The dstrbuton of RSS wth the user n Fgure 2-a s also skewed to the left. The leftskew property seems to occur n most of our measurements as reported by the skewness n Tables, 2, and 3. Ths property s usually observed when the data have an upper bound whch s the case for the attenuated RSS measurement. In a comparson of the 46 dstrbutons, we saw that 39 of them had dstrbutons that were left-skewed whle fve of them were almost symmetrc and only two of them were rght-skewed. The left-skew s the effect of the range lmtaton mposed by the maxmum RSS at each locaton. Because of the complexty of the rado propagaton, the dstrbuton of RSS s dffcult to model and ft to well-known dstrbutons. The authors n [6] conclude that they rather record the dstrbuton of the RSS than reduce t to only the mean value. We

5 beleve that t would be a great beneft f we could fnd a representatve or approxmate dstrbuton of the underlyng RSS process for understandng locaton fngerprntng. Ths s part of our ongong research The standard devaton of the RSS. The results n Table, 2, and 3 also reveal an nterestng trend n the second order statstcs of the RSS values. The standard devatons are qute smlar for a sgnal from the same AP at a partcular locaton except when the user s orentaton blocks that AP. The man dfference between the three APs s the dstance to the measurement pont L2. Comparng the three tables, the results ndcate that the farther the AP s from the MS or the lower the receved sgnal level s, the smaller the degree of standard devaton. Note that wthn the same table the RSS data whch s blocked by the user has a smaller standard devaton. Ths occurs as a result of the sgnal attenuaton by the user s body and the smaller range of RSS between the maxmum and mnmum recevable sgnal level. A lnear regresson plot (Fgure 3) between the sample mean of the RSS and the standard devaton collected from 46 dstrbutons llustrates the trend. Ths observaton suggests that the RSS values from the same AP at two dfferent locatons may be dffcult to dstngush for postonng purposes when the RSS level s hgh n whch case t tends to have a large degree of varaton. A good communcatons sgnal may not result n a good postonng sgnal. On the other hand, two nearby locatons mght be easly dentfed f both have low RSS levels and smaller sgnal varatons. Ths good postonng case usually occurs n ndoor locatons wth NLOS. Ths s rather counter-ntutve snce the farther apart the WLAN recever s from the AP, the worse the measurement accuracy or the larger the sgnal varaton should be as suggested by [5]. Sample Standard Devaton n dbm Sample Mean RSS Lnear Regresson s Y =.497 X Sample Mean of RSS n dbm Fgure 3. The relatonshp between average RSS and ts standard devaton Statonarty of the RSS. Assumng that the ergodc theorem s appled accordng to the Wener defnton of statonarty [], we analyze the statonarty of the RSS element by breakng the seres of RSS measurements nto separate peces over dfferent tme ntervals. A random process s sad to be statonary when t meets two condtons. Frst, ts mean and varance reman the same over tme. Second, ts autocovarance functon has the same shape for each separate tme-seres. We nvestgated ths property over two tme scales: peces of 5 mnutes wthn the same hour and peces of one hour over fve dfferent hours. After dvdng the seres of measurement data of RSS n Fgure 2-a wthn the same hour nto groups of 5 mnutes, the RSS dstrbuton wthn each quarter s observed to follow a smlar dstrbuton wthn the same group. Table 4 lsts the summary statstcs wthn each quarter. These results suggest that the RSS dstrbuton may be statonary or tme ndependent snce the means and the sample varances of each quarter are very close together. The correlograms n Fgure 4 depct the same shapes for each quarter ndcatng that the second condton s also met for ths tme scale. Autocorrelaton Table 4. Mean and standard devaton of RSS wth user (dbm) Statstcs st Qtr. 2 nd Qtr. 3 rd Qtr. 4 th Qtr. Mean Standard Devaton Sample Varance Correlogram of st Quarter Tme lag n second Correlogram of 3rd Quarter Autocorrelaton Tme lag n second Autocorrelaton Correlogram of 2nd Quarter Tme lag n second Correlogram of 4th Quarter Autocorrelaton Tme lag n second Fgure 4. The correlograms of RSS wthn the same hour

6 The one-hour tme scale study was made for a sgnal measured at L over dfferent tmes of day. Table 5 shows a consstent mean, but nconsstent varance values of the RSS. Therefore, the test for the frst condton for statonarty fals and we conclude that the RSS random process n ths case s nonstatonary. Fgure 5 llustrates sample paths of average receved sgnal strength values from the three access ponts measured at locaton L3 over a perod of one hour. Observe that the sample mean from SIS4 abruptly changes to another value (-7 to -6 dbm) whch confrms our concluson on the non-statonary property. Ths s common due to the changng ndoor envronment such as n ths case when a person walked nto the room and sat n the mddle of the room IS 4 after approxmately 3 mnutes nto our experment. Notce that the rest of the receved sgnals from other APs located outsde the room are not affected by ths event. Although the statonary assumpton may not be vald over all tme scales, there s some evdence that we could assume statonarty over small tme scale for modelng purposes. Table 5. Tme dependency of RSS (dbm) from SIS4 wth user s presence Statstcs AM 2AM 2PM 8PM PM Mean Standard Devaton Sample Varance Average receved sgnal strength n dbm AP (ss4) AP2 (ss4) AP3 (ss5) Tme n second Fgure 5. Samples of RSS from three APs Tme dependency of receved sgnal strength. The summary of statstcs n Table 5 suggests that there s some dependency of RSS on the tme of day. Ths s due to the dynamcs of the ndoor envronment. To understand the property better, we requre more measurements and ths s part of our ongong research work Propertes of multple RSSs at a partcular locaton Ths subsecton analyzes the dependency of multple RSSs from multple APs. Ths s to confrm an ntuton that the RSS from multple APs are actually ndependent. The second part of ths subsecton dscusses the effect of nterference on the RSS when there s another AP transmttng n the same frequency channel Independence of multple RSSs. The average of the RSS from each AP s a value of a locaton fngerprnt vector. To verfy statstcal ndependence between these values, a measurement of multple RSS samples was collected at locaton 3 (L3) n Fgure where the MS can receve sgnals from three APs smultaneously durng an afternoon hour. The dstances from the three APs SIS4, SIS4, and SIS5 were approxmately 8, 5, and meters. We took RSS measurements for approxmately one hour wth the user s presence and the sample paths of RSS levels are showed n Fgure 5. The standard devatons from each AP were 5.67, 2.23, and.8, respectvely. The means were , -8.23, and dbm. The correlaton values between each par of RSS data are C (SIS4,SIS4) = -.2, C (SIS4,SIS5) =.3, C (SIS4,SIS5) = -.3. Therefore, we can conclude that the RSS from the APs are uncorrelated Interference from multple APs. There are two APs n our experment that use the same frequency channel number 6, whch are SIS4 and SIS5. One may thnk that the RSSs from both APs mght nterfere wth each other and cause dffculty n formng the locaton fngerprntng. However, our ntal results as calculated by the correlaton ndcate that both RSSs are ndependent and do not nterfere wth the recepton of each other. Ths s due to the way n whch the 82. MAC operates where a transmsson s ether not heard or s deferred f a competng transmsson exsts Propertes of RSS at dfferent locatons Generally, the RSS falls lnearly wth the log of the dstance between the transmtter and the recever. Ths subsecton wll not focus on ths common knowledge but wll nstead nvestgate how the samples of RSS fngerprnts at dfferent locatons may look lke. We plot the RSS patterns n two dmensons n order to analyze the clusterng of the RSS data. The clusterng of two RSS data at two locatons s plotted n Fgure 6.

7 RSS of SIS4 n dbm Locaton 2 Locaton RSS of SIS4 n dbm Fgure 6. RSS fngerprnts wth two elements The plot n Fgure 6 contans all possble vectors P = [ 2 ] for L2 and L3 wth on the x-axs and 2 on the y-axs. We see that the locaton fngerprnts can be separated by some dscrmnant functon or clusterng technque. Note that L3 conssts of 3,666 samples and L2 conssts of 3,465 samples. Only certan patterns are present whch mples that there are fewer unque fngerprnts for each locaton. Fgure 7 shows the densty of each fngerprnt. Ths vsual study suggests that we may use the center of the cluster as a representaton of the locaton fngerprnt nstead of the dstrbuton tself as these locatons can be clearly separated. Fgure 7 suggests that only two APs are suffcent to dstngush between locatons for a system wth small number of postons and coarse locaton granularty. Two locatons become dffcult to dentfy f ther patterns are closer together and exhbt large varatons due to the nature of standard devaton of the RSS. Increasng the number of APs s one way to further separate two locaton fngerprnts. Frequency ss4 (dbm) L L2 6 5 ss4 (dbm) Fgure 7. Frequency of occurrence of RSS patterns Implcatons on postonng algorthms The data analyses n the prevous secton have two major mplcatons on selectng potental postonng algorthms. Frst, we observed that the RSS sample vectors exhbt clusterng (they are concentrated around the center of a cluster wth the greatest frequency of occurrence beng roughly at the center as n Fgure 7). Ths explans why the accuracy performance of all postonng algorthms evaluated n [2] s smlar. Both the support vector machnes (SVMs) approach and the k-nearest neghbors (or equvalently Eucldean dstance) approach provde smlar accuracy and precson. The reason s that locaton fngerprnts can be smply represented by vectors of average RSSs. Second, to perform any analytcal performance evaluaton of pattern classfers, we eventually need a model of locaton fngerprnts. In the followng secton we suggest a smple model of locaton fngerprnts and compare t wth the emprcal ones from ths study based on our prevous work n [3]. 4. Modelng of RSS for locaton fngerprnts and performance of the Eucldean dstance Although our prelmnary fndng suggests that the RSS dstrbuton s not normal, some of them could be approxmated by a normal dstrbuton. In an attempt to model ndoor postonng systems [3], we assumed a frst cut Gaussan model of the RSS. A locaton fngerprnt s denoted as a vector of N features of RSS, e.g. X = (x, x 2,, x 3 ). Here, we lst the assumptons made n [3] that are partally supported by the work here. Each RSS feature n fngerprnt s normally dstrbuted and statonary over a small tme scale. The sample standard devatons of all RSS features are assumed to be constant and unque for each RSS. (We dd not model the varaton n standard devaton for each RSS feature.) The mean of the RSS can be used as the fngerprnt as samples of the RSS vector exhbt clusterng. All RSS features n each locaton fngerprnt are mutually ndependent (whch s confrmed by our measurement n Secton 3.3). We compare the locaton estmaton performance usng ths smplfed model wth the one usng emprcal dstrbutons obtaned from the measurement results of ths paper. The smple locaton estmaton algorthm utlzes the Eucldean dstance. The number of access ponts N s three. The mathematcal

8 expresson for the accuracy was developed n [3]. We use the expressons to predct the performance of two smplfed postonng systems n ths secton. 4. Two-locaton system As an example of a system wth only two locaton fngerprnts, we select the measurement data of locatons L2 and L3 for analyss. Note that these are ncely dstngushable. Let R = (r, r 2,, r 3 ) and S = (s, s 2,, s 3 ) be the locaton fngerprnts (mean values of the RSSs from the APs) of L2 and L3, respectvely. The actual physcal dstance between the two locatons s 8 ft (5.5 m). Table 6 summarzes the locaton fngerprnts and ther standard devatons. Table 6. Fngerprnts of two-poston system (dbm). Locaton Statstcs SIS4 SIS4 SIS5 L2 Mean STD L3 Mean STD Usng a Monte Carlo smulaton to generate the samples of locaton fngerprnts based on the emprcal dstrbutons, we compare the error of locaton detecton gven that L2 s the correct locaton wth the analytcal mathematcal formulaton n [3] and smulatons wth a Gaussan dstrbuton wth the same means and standard devatons. The probablty of ncorrectly pckng locaton L3 nstead of L2 s Pe. For a system wth two postons, Pe can be found from the probablty of returnng the correct locaton (Pc) as: c Pe Pc erf, () c N 2 c 2r c 2 2 s r, and r s N where,, N ( 2 ) 2. We performed 2 smulatons of, samples each. The results are summarzed n Table 7. The Gaussan approxmaton provdes an optmstc error performance for the system. The analytcal result from () exactly calculates Pe. The emprcal dstrbutons have a worse performance because there are some samples of RSS fngerprnts that are closer to the locaton L3 s fngerprnt than L2 s fngerprnt. Fgure 6, whch represents the projecton of all fngerprnts nto the plane of two features, clearly explans the cause of the worse performance because there are some RSS patterns on the left-sde of the plot that may be wrongly nterpreted as correspondng to L3. Table 7 also shows the results when we consder only two APs (SIS4, SIS5) and one AP (SIS4) n whch the error probabltes are hgher. Table 7. Probablty of returnng ncorrect locaton at 95% C.I. for two-locaton system. Scenaro 3 APs 2 APs AP Emprcal.22e e-5.79e-3 3.2e-5 6.9e e-5 Gaussan.24e e-6.4e e e-3 4.6e-5 Analytcal.24e-3.4e-3 7.5e Twenty fve-locaton system The second system we consder conssts of 25 locaton fngerprnts where we measured the real locaton fngerprnts from a grd of 25 postons. The grd spacng s approxmately meter. The locaton area of 6 m 2 covers part of the room IS4 and part of the corrdor as showed n Fgure 8. Each poston s labeled by the square box North Soft Wall Room IS Fgure 8. Grd of system wth 25 locatons Due to space lmtaton, we wll not lst the measured locaton fngerprnts here. Note that the measurement s done for only one orentaton of the user (facng north). Each poston can receve sgnals clearly from all three APs (SIS4, SIS4, SIS5) and there are approxmately,2 samples of each RSS measured over fve mnutes at a rate of four samples per second. We assume that the current MS s locaton s at the center of the map. A smulaton was done n the same manner as the two locaton system. The results are summarzed n Table 8. In order to fnd the exact analytcal expresson, we need to fnd the jont probablty densty of multple locaton fngerprnts whch s rather cumbersome. We argue n [3] that ths can be avoded by resortng to an approxmaton

9 of the probablty of returnng an erroneous locaton. Ths consders only the locaton fngerprnts of M N = 8 nearest neghborng postons n two-dmensons and the correct locaton. The approxmaton works for a large number of access ponts and can be wrtten as: Pe Pc j. (2) jm N Table 8. Probablty of returnng ncorrect locaton at 9% C.I. for 25 locatons. Scenaro 3 APs 2 APs AP Emprcal.67.46e e e-3 Gaussan.85.67e e e-3 Approxmaton The overall performance results n ths system are worse than the prevous one because there are more locatons to be compared wth and the spacng between the locatons s much closer. Dependng on the standard devaton of each RSS feature, t s possble that the fngerprnt patterns overlap. The results wth a Gaussan model are worse than those wth the emprcal model and appear to be rather pessmstc n ths case. Ths can be explaned as follows: the Gaussan model tends to spread the fngerprnt patterns evenly around the mean value whle the real fngerprnt patterns are asymmetrc and more concentrated around ther means n some cases. The analytcal approxmaton s hghly pessmstc and not approprate n ths case. In summary the Gaussan model provdes a probablty of an erroneous report on the same order as one would expect wth real RSS samples, but could be ether optmstc or pessmstc. 5. Conclusons We present an ntal analyss of the RSS values reported by an 82.b NIC commonly used n ndoor locaton systems based on locaton fngerprntng. We pont out that the user s presence should be taken nto account when collectng the locaton fngerprnt for user related locaton-based servces. The effect of user s orentaton s sgnfcant and the orentaton should be recorded n the database as demonstrated n []. We also analyze the statstcal propertes of the RSS and we fnd that t s statonary under certan crcumstances, but n general, such a concluson cannot be made. The dstrbuton of the RSS s not usually Gaussan, t s often left-skewed and the standard devaton vares accordng to the sgnal level. It s clear from our measurements that sgnals from multple APs are mostly ndependent and the nterference from other APs usng the same frequency does not have a sgnfcant mpact on the RSS pattern. The vsual presentatons of the RSS patterns n Secton 3.4 show that the fngerprnt can be grouped together as a set of clusters. More than one cluster may represent one locaton because of the multmodal dstrbuton of the RSS. In such a case, usng a smple Eucldean dstance as n [] to determne the locaton may classfy some patterns nto a wrong locaton easly. Ths causes poor performance of a postonng system that uses the Eucldean dstance. Fnally, n the last secton, we compare the error performance of a poston locaton system wth real locaton fngerprnts and a frst cut Gaussan model. The results ndcate that our model of locaton fngerprnts provde some approxmatons of the performance. The Gaussan model s ether optmstc or pessmstc but provdes values on the same order as a real system. The future work s to look for alternatve models for the dstrbuton of the RSS and to understand how they mpact poston locaton further. The results n ths paper and our prevous work could provde nsght on the mechansm behnd ndoor poston locaton systems based on locaton fngerprntng. In concluson, ths study steps back and nvestgates the RSS pattern n a greater detal. Ths study rases an mportant aspect of desgnng a locaton fngerprntng system that s to examne the propertes of the fngerprnt tself before applyng a pattern recognton technque to solve the postonng problem. Smple pattern recognton technques may be sutable and more effcent than more sophstcated ones. More extensve measurement campagns are needed to verfy some of the propertes lsted here. However, the lessons learned from ths study are used to support a theoretcal model of locaton fngerprnt whch n turn can be used to create a desgn framework of an ndoor postonng system [3]. Gven a framework and theoretcal explanaton, any future study and desgn of locaton fngerprntng system should be more effcent and less tme consumng. We could reduce the measurement tme n order to fne tune the system performance such as accuracy and precson to meet any requred crtera. 6. References [] P. Bahl and V. N. Padmanabhan, RADAR: An In- Buldng RF-based User Locaton and Trackng System, n Proc. IEEE INFOCOM, 2, pp [2] G. M. Djuknc, and R. E. Rchton, Geolocaton and Asssted GPS, IEEE Computer, vol. 2, pp , Feb. 2. [3] M. A. Youssef, A. Agrawala, and A. U. Shankar, WLAN Locaton Determnaton va Clusterng and

10 Probablty Dstrbutons, n Proc. IEEE PerCom, Mar. 23. [4] T. Roos et al., A Probablstc Approach to WLAN User Locaton, Int l Journal of Wreless Informaton Networks, vol. 9, pp , Jul. 22. [5] J. Small, A. Smalagc, and D. P. Seworek, (2, Dec.) Determnng User Locaton For Context Aware Computng Through the Use of a Wreless LAN Infrastructure. [Onlne at] aura/docdr/small.pdf. [6] A. M. Ladd et. al., Robotcs-Based Locaton Sensng usng Wreless Ethernet, n Proc. MOBICOM, 22, pp [7] The Insttute of Electrcal and Electroncs Engneers, Inc. IEEE Standards Wreless LAN Medum Access Control (MAC) and Physcal Layer (PHY) specfcatons [8] Agere System Inc. (23, Feb.) Agere Product Bref: WaveLAN 82.b chp set. [Onlne] Avalable: [9] K. Pahlavan and P. Krshnamurthy, Prncples of Wreless Networks: A Unfed Approach, Prentce Hall PTR, Upper Saddle Rver, New Jersey, 22. [] B. Sklar, Raylegh fadng channels n moble dgtal communcaton systems: I. Characterzaton, IEEE Communcatons Magazne, vol. 35, pp. 9-, Jul [] J. M. Gottman, Tme-seres Analyss: A Comprehensve Introducton for Socal Scentsts, Cambrdge Unversty Press, New York, NY, 98. [2] R. Battt, M. Brunato, and A. Vllan. (22, Oct.) Statstcal Learnng Theory for Locaton Fngerprntng n Wreless LANs. Techncal Report DIT-2-86, Unverst a d Trento. [Onlne] Avalable: ~battt/archve/86.pdf [3] K. Kaemarungs, and P. Krshnamurthy, Modelng of Indoor Postonng Systems Based on Locaton Fngerprntng, n Proc. IEEE INFOCOM, May 24.

INVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS

INVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS 21 22 September 2007, BULGARIA 119 Proceedngs of the Internatonal Conference on Informaton Technologes (InfoTech-2007) 21 st 22 nd September 2007, Bulgara vol. 2 INVESTIGATION OF VEHICULAR USERS FAIRNESS

More information

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ). REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or

More information

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.

More information

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol

CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK. Sample Stability Protocol CHOLESTEROL REFERENCE METHOD LABORATORY NETWORK Sample Stablty Protocol Background The Cholesterol Reference Method Laboratory Network (CRMLN) developed certfcaton protocols for total cholesterol, HDL

More information

An Empirical Study of Search Engine Advertising Effectiveness

An Empirical Study of Search Engine Advertising Effectiveness An Emprcal Study of Search Engne Advertsng Effectveness Sanjog Msra, Smon School of Busness Unversty of Rochester Edeal Pnker, Smon School of Busness Unversty of Rochester Alan Rmm-Kaufman, Rmm-Kaufman

More information

What is Candidate Sampling

What is Candidate Sampling What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble

More information

An Interest-Oriented Network Evolution Mechanism for Online Communities

An Interest-Oriented Network Evolution Mechanism for Online Communities An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne

More information

Can Auto Liability Insurance Purchases Signal Risk Attitude?

Can Auto Liability Insurance Purchases Signal Risk Attitude? Internatonal Journal of Busness and Economcs, 2011, Vol. 10, No. 2, 159-164 Can Auto Lablty Insurance Purchases Sgnal Rsk Atttude? Chu-Shu L Department of Internatonal Busness, Asa Unversty, Tawan Sheng-Chang

More information

An Alternative Way to Measure Private Equity Performance

An Alternative Way to Measure Private Equity Performance An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate

More information

Calculating the high frequency transmission line parameters of power cables

Calculating the high frequency transmission line parameters of power cables < ' Calculatng the hgh frequency transmsson lne parameters of power cables Authors: Dr. John Dcknson, Laboratory Servces Manager, N 0 RW E B Communcatons Mr. Peter J. Ncholson, Project Assgnment Manager,

More information

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..

More information

Forecasting the Direction and Strength of Stock Market Movement

Forecasting the Direction and Strength of Stock Market Movement Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems

More information

DEFINING %COMPLETE IN MICROSOFT PROJECT

DEFINING %COMPLETE IN MICROSOFT PROJECT CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,

More information

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services An Evaluaton of the Extended Logstc, Smple Logstc, and Gompertz Models for Forecastng Short Lfecycle Products and Servces Charles V. Trappey a,1, Hsn-yng Wu b a Professor (Management Scence), Natonal Chao

More information

Causal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting

Causal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting Causal, Explanatory Forecastng Assumes cause-and-effect relatonshp between system nputs and ts output Forecastng wth Regresson Analyss Rchard S. Barr Inputs System Cause + Effect Relatonshp The job of

More information

The OC Curve of Attribute Acceptance Plans

The OC Curve of Attribute Acceptance Plans The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4

More information

An Introduction to 3G Monte-Carlo simulations within ProMan

An Introduction to 3G Monte-Carlo simulations within ProMan An Introducton to 3G Monte-Carlo smulatons wthn ProMan responsble edtor: Hermann Buddendck AWE Communcatons GmbH Otto-Llenthal-Str. 36 D-71034 Böblngen Phone: +49 70 31 71 49 7-16 Fax: +49 70 31 71 49

More information

On-Line Fault Detection in Wind Turbine Transmission System using Adaptive Filter and Robust Statistical Features

On-Line Fault Detection in Wind Turbine Transmission System using Adaptive Filter and Robust Statistical Features On-Lne Fault Detecton n Wnd Turbne Transmsson System usng Adaptve Flter and Robust Statstcal Features Ruoyu L Remote Dagnostcs Center SKF USA Inc. 3443 N. Sam Houston Pkwy., Houston TX 77086 Emal: ruoyu.l@skf.com

More information

Performance Analysis of Energy Consumption of Smartphone Running Mobile Hotspot Application

Performance Analysis of Energy Consumption of Smartphone Running Mobile Hotspot Application Internatonal Journal of mart Grd and lean Energy Performance Analyss of Energy onsumpton of martphone Runnng Moble Hotspot Applcaton Yun on hung a chool of Electronc Engneerng, oongsl Unversty, 511 angdo-dong,

More information

Recurrence. 1 Definitions and main statements

Recurrence. 1 Definitions and main statements Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.

More information

BERNSTEIN POLYNOMIALS

BERNSTEIN POLYNOMIALS On-Lne Geometrc Modelng Notes BERNSTEIN POLYNOMIALS Kenneth I. Joy Vsualzaton and Graphcs Research Group Department of Computer Scence Unversty of Calforna, Davs Overvew Polynomals are ncredbly useful

More information

VoIP Playout Buffer Adjustment using Adaptive Estimation of Network Delays

VoIP Playout Buffer Adjustment using Adaptive Estimation of Network Delays VoIP Playout Buffer Adjustment usng Adaptve Estmaton of Network Delays Mroslaw Narbutt and Lam Murphy* Department of Computer Scence Unversty College Dubln, Belfeld, Dubln, IRELAND Abstract The poor qualty

More information

Forecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network

Forecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network 700 Proceedngs of the 8th Internatonal Conference on Innovaton & Management Forecastng the Demand of Emergency Supples: Based on the CBR Theory and BP Neural Network Fu Deqang, Lu Yun, L Changbng School

More information

Luby s Alg. for Maximal Independent Sets using Pairwise Independence

Luby s Alg. for Maximal Independent Sets using Pairwise Independence Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent

More information

Characterization of Assembly. Variation Analysis Methods. A Thesis. Presented to the. Department of Mechanical Engineering. Brigham Young University

Characterization of Assembly. Variation Analysis Methods. A Thesis. Presented to the. Department of Mechanical Engineering. Brigham Young University Characterzaton of Assembly Varaton Analyss Methods A Thess Presented to the Department of Mechancal Engneerng Brgham Young Unversty In Partal Fulfllment of the Requrements for the Degree Master of Scence

More information

Face Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching)

Face Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching) Face Recognton Problem Face Verfcaton Problem Face Verfcaton (1:1 matchng) Querymage face query Face Recognton (1:N matchng) database Applcaton: Access Control www.vsage.com www.vsoncs.com Bometrc Authentcaton

More information

Conversion between the vector and raster data structures using Fuzzy Geographical Entities

Conversion between the vector and raster data structures using Fuzzy Geographical Entities Converson between the vector and raster data structures usng Fuzzy Geographcal Enttes Cdála Fonte Department of Mathematcs Faculty of Scences and Technology Unversty of Combra, Apartado 38, 3 454 Combra,

More information

Damage detection in composite laminates using coin-tap method

Damage detection in composite laminates using coin-tap method Damage detecton n composte lamnates usng con-tap method S.J. Km Korea Aerospace Research Insttute, 45 Eoeun-Dong, Youseong-Gu, 35-333 Daejeon, Republc of Korea yaeln@kar.re.kr 45 The con-tap test has the

More information

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek HE DISRIBUION OF LOAN PORFOLIO VALUE * Oldrch Alfons Vascek he amount of captal necessary to support a portfolo of debt securtes depends on the probablty dstrbuton of the portfolo loss. Consder a portfolo

More information

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 12

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 12 14 The Ch-squared dstrbuton PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 1 If a normal varable X, havng mean µ and varance σ, s standardsed, the new varable Z has a mean 0 and varance 1. When ths standardsed

More information

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT Toshhko Oda (1), Kochro Iwaoka (2) (1), (2) Infrastructure Systems Busness Unt, Panasonc System Networks Co., Ltd. Saedo-cho

More information

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION Vson Mouse Saurabh Sarkar a* a Unversty of Cncnnat, Cncnnat, USA ABSTRACT The report dscusses a vson based approach towards trackng of eyes and fngers. The report descrbes the process of locatng the possble

More information

"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *

Research Note APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES * Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC

More information

CHAPTER 14 MORE ABOUT REGRESSION

CHAPTER 14 MORE ABOUT REGRESSION CHAPTER 14 MORE ABOUT REGRESSION We learned n Chapter 5 that often a straght lne descrbes the pattern of a relatonshp between two quanttatve varables. For nstance, n Example 5.1 we explored the relatonshp

More information

OPT Online Person Tracking System for Context-awareness in Wireless Personal Network

OPT Online Person Tracking System for Context-awareness in Wireless Personal Network OPT Onlne Person Trackng System for Context-awareness n Wreless Personal Network Xuel An R. Venkatesha Prasad Jng Wang I.G.M.M. Nemegeers EEMCS, Delft Unversty of Technology, The Netherlands {x.an, j.wang3,vprasad,gnas.n}@ew.tudelft.nl

More information

Statistical Methods to Develop Rating Models

Statistical Methods to Develop Rating Models Statstcal Methods to Develop Ratng Models [Evelyn Hayden and Danel Porath, Österrechsche Natonalbank and Unversty of Appled Scences at Manz] Source: The Basel II Rsk Parameters Estmaton, Valdaton, and

More information

A Revised Received Signal Strength Based Localization for Healthcare

A Revised Received Signal Strength Based Localization for Healthcare , pp.273-282 http://dx.do.org/10.14257/mue.2015.10.10.27 A Revsed Receved Sgnal Strength Based Localzaton for Healthcare Wenhuan Ch 1, Yuan Tan 2, Mznah Al-Rodhaan 2, Abdullah Al-Dhelaan 2 and Yuanfeng

More information

How To Calculate The Accountng Perod Of Nequalty

How To Calculate The Accountng Perod Of Nequalty Inequalty and The Accountng Perod Quentn Wodon and Shlomo Ytzha World Ban and Hebrew Unversty September Abstract Income nequalty typcally declnes wth the length of tme taen nto account for measurement.

More information

Realistic Image Synthesis

Realistic Image Synthesis Realstc Image Synthess - Combned Samplng and Path Tracng - Phlpp Slusallek Karol Myszkowsk Vncent Pegoraro Overvew: Today Combned Samplng (Multple Importance Samplng) Renderng and Measurng Equaton Random

More information

How To Understand The Results Of The German Meris Cloud And Water Vapour Product

How To Understand The Results Of The German Meris Cloud And Water Vapour Product Ttel: Project: Doc. No.: MERIS level 3 cloud and water vapour products MAPP MAPP-ATBD-ClWVL3 Issue: 1 Revson: 0 Date: 9.12.1998 Functon Name Organsaton Sgnature Date Author: Bennartz FUB Preusker FUB Schüller

More information

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP)

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP) 6.3 / -- Communcaton Networks II (Görg) SS20 -- www.comnets.un-bremen.de Communcaton Networks II Contents. Fundamentals of probablty theory 2. Emergence of communcaton traffc 3. Stochastc & Markovan Processes

More information

Traffic State Estimation in the Traffic Management Center of Berlin

Traffic State Estimation in the Traffic Management Center of Berlin Traffc State Estmaton n the Traffc Management Center of Berln Authors: Peter Vortsch, PTV AG, Stumpfstrasse, D-763 Karlsruhe, Germany phone ++49/72/965/35, emal peter.vortsch@ptv.de Peter Möhl, PTV AG,

More information

Brigid Mullany, Ph.D University of North Carolina, Charlotte

Brigid Mullany, Ph.D University of North Carolina, Charlotte Evaluaton And Comparson Of The Dfferent Standards Used To Defne The Postonal Accuracy And Repeatablty Of Numercally Controlled Machnng Center Axes Brgd Mullany, Ph.D Unversty of North Carolna, Charlotte

More information

IMPACT ANALYSIS OF A CELLULAR PHONE

IMPACT ANALYSIS OF A CELLULAR PHONE 4 th ASA & μeta Internatonal Conference IMPACT AALYSIS OF A CELLULAR PHOE We Lu, 2 Hongy L Bejng FEAonlne Engneerng Co.,Ltd. Bejng, Chna ABSTRACT Drop test smulaton plays an mportant role n nvestgatng

More information

Cooperative Load Balancing in IEEE 802.11 Networks with Cell Breathing

Cooperative Load Balancing in IEEE 802.11 Networks with Cell Breathing Cooperatve Load Balancng n IEEE 82.11 Networks wth Cell Breathng Eduard Garca Rafael Vdal Josep Paradells Wreless Networks Group - Techncal Unversty of Catalona (UPC) {eduardg, rvdal, teljpa}@entel.upc.edu;

More information

Risk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008

Risk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008 Rsk-based Fatgue Estmate of Deep Water Rsers -- Course Project for EM388F: Fracture Mechancs, Sprng 2008 Chen Sh Department of Cvl, Archtectural, and Envronmental Engneerng The Unversty of Texas at Austn

More information

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network *

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 819-840 (2008) Data Broadcast on a Mult-System Heterogeneous Overlayed Wreless Network * Department of Computer Scence Natonal Chao Tung Unversty Hsnchu,

More information

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by 6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng

More information

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo.

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo. ICSV4 Carns Australa 9- July, 007 RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL Yaoq FENG, Hanpng QIU Dynamc Test Laboratory, BISEE Chna Academy of Space Technology (CAST) yaoq.feng@yahoo.com Abstract

More information

Scale Dependence of Overconfidence in Stock Market Volatility Forecasts

Scale Dependence of Overconfidence in Stock Market Volatility Forecasts Scale Dependence of Overconfdence n Stoc Maret Volatlty Forecasts Marus Glaser, Thomas Langer, Jens Reynders, Martn Weber* June 7, 007 Abstract In ths study, we analyze whether volatlty forecasts (judgmental

More information

Performance Analysis of Time-of-Arrival Mobile Positioning in Wireless Cellular CDMA Networks

Performance Analysis of Time-of-Arrival Mobile Positioning in Wireless Cellular CDMA Networks Performance Analyss of Tme-of-Arrval Moble Postonng n Wreless Cellular CDMA Networks 437 1 X Performance Analyss of Tme-of-Arrval Moble Postonng n Wreless Cellular CDMA Networks M. A.Landols, A. H. Muqabel,

More information

A Hierarchical Anomaly Network Intrusion Detection System using Neural Network Classification

A Hierarchical Anomaly Network Intrusion Detection System using Neural Network Classification IDC IDC A Herarchcal Anomaly Network Intruson Detecton System usng Neural Network Classfcaton ZHENG ZHANG, JUN LI, C. N. MANIKOPOULOS, JAY JORGENSON and JOSE UCLES ECE Department, New Jersey Inst. of Tech.,

More information

On the Optimal Control of a Cascade of Hydro-Electric Power Stations

On the Optimal Control of a Cascade of Hydro-Electric Power Stations On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;

More information

MONITORING METHODOLOGY TO ASSESS THE PERFORMANCE OF GSM NETWORKS

MONITORING METHODOLOGY TO ASSESS THE PERFORMANCE OF GSM NETWORKS Electronc Communcatons Commttee (ECC) wthn the European Conference of Postal and Telecommuncatons Admnstratons (CEPT) MONITORING METHODOLOGY TO ASSESS THE PERFORMANCE OF GSM NETWORKS Athens, February 2008

More information

Analysis of Energy-Conserving Access Protocols for Wireless Identification Networks

Analysis of Energy-Conserving Access Protocols for Wireless Identification Networks From the Proceedngs of Internatonal Conference on Telecommuncaton Systems (ITC-97), March 2-23, 1997. 1 Analyss of Energy-Conservng Access Protocols for Wreless Identfcaton etworks Imrch Chlamtac a, Chara

More information

VOLUME 5 BLAGOEVGRAD, BULGARIA SCIENTIFIC. Research ELECTRONIC ISSUE ISSN 1312-7535

VOLUME 5 BLAGOEVGRAD, BULGARIA SCIENTIFIC. Research ELECTRONIC ISSUE ISSN 1312-7535 VOLUME 5 007 BLAGOEVGRAD, BULGARIA SCIENTIFIC Research ISSN 131-7535 ELECTRONIC ISSUE IMPROVING FAIRNESS IN CDMA-HDR NETWORKS Valentn Hrstov Abstract. Improvng throughput and farness n Cellular Data Networks

More information

Exhaustive Regression. An Exploration of Regression-Based Data Mining Techniques Using Super Computation

Exhaustive Regression. An Exploration of Regression-Based Data Mining Techniques Using Super Computation Exhaustve Regresson An Exploraton of Regresson-Based Data Mnng Technques Usng Super Computaton Antony Daves, Ph.D. Assocate Professor of Economcs Duquesne Unversty Pttsburgh, PA 58 Research Fellow The

More information

Risk Model of Long-Term Production Scheduling in Open Pit Gold Mining

Risk Model of Long-Term Production Scheduling in Open Pit Gold Mining Rsk Model of Long-Term Producton Schedulng n Open Pt Gold Mnng R Halatchev 1 and P Lever 2 ABSTRACT Open pt gold mnng s an mportant sector of the Australan mnng ndustry. It uses large amounts of nvestments,

More information

L10: Linear discriminants analysis

L10: Linear discriminants analysis L0: Lnear dscrmnants analyss Lnear dscrmnant analyss, two classes Lnear dscrmnant analyss, C classes LDA vs. PCA Lmtatons of LDA Varants of LDA Other dmensonalty reducton methods CSCE 666 Pattern Analyss

More information

The Application of Fractional Brownian Motion in Option Pricing

The Application of Fractional Brownian Motion in Option Pricing Vol. 0, No. (05), pp. 73-8 http://dx.do.org/0.457/jmue.05.0..6 The Applcaton of Fractonal Brownan Moton n Opton Prcng Qng-xn Zhou School of Basc Scence,arbn Unversty of Commerce,arbn zhouqngxn98@6.com

More information

Time Domain simulation of PD Propagation in XLPE Cables Considering Frequency Dependent Parameters

Time Domain simulation of PD Propagation in XLPE Cables Considering Frequency Dependent Parameters Internatonal Journal of Smart Grd and Clean Energy Tme Doman smulaton of PD Propagaton n XLPE Cables Consderng Frequency Dependent Parameters We Zhang a, Jan He b, Ln Tan b, Xuejun Lv b, Hong-Je L a *

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan support vector machnes.

More information

ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING

ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING ANALYZING THE RELATIONSHIPS BETWEEN QUALITY, TIME, AND COST IN PROJECT MANAGEMENT DECISION MAKING Matthew J. Lberatore, Department of Management and Operatons, Vllanova Unversty, Vllanova, PA 19085, 610-519-4390,

More information

8 Algorithm for Binary Searching in Trees

8 Algorithm for Binary Searching in Trees 8 Algorthm for Bnary Searchng n Trees In ths secton we present our algorthm for bnary searchng n trees. A crucal observaton employed by the algorthm s that ths problem can be effcently solved when the

More information

denote the location of a node, and suppose node X . This transmission causes a successful reception by node X for any other node

denote the location of a node, and suppose node X . This transmission causes a successful reception by node X for any other node Fnal Report of EE359 Class Proect Throughput and Delay n Wreless Ad Hoc Networs Changhua He changhua@stanford.edu Abstract: Networ throughput and pacet delay are the two most mportant parameters to evaluate

More information

FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES

FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES Zuzanna BRO EK-MUCHA, Grzegorz ZADORA, 2 Insttute of Forensc Research, Cracow, Poland 2 Faculty of Chemstry, Jagellonan

More information

Slow Fading Channel Selection: A Restless Multi-Armed Bandit Formulation

Slow Fading Channel Selection: A Restless Multi-Armed Bandit Formulation Slow Fadng Channel Selecton: A Restless Mult-Armed Bandt Formulaton Konstantn Avrachenkov INRIA, Maestro Team BP95, 06902 Sopha Antpols, France Emal: k.avrachenkov@sopha.nra.fr Laura Cottatellucc, Lorenzo

More information

RELIABILITY, RISK AND AVAILABILITY ANLYSIS OF A CONTAINER GANTRY CRANE ABSTRACT

RELIABILITY, RISK AND AVAILABILITY ANLYSIS OF A CONTAINER GANTRY CRANE ABSTRACT Kolowrock Krzysztof Joanna oszynska MODELLING ENVIRONMENT AND INFRATRUCTURE INFLUENCE ON RELIABILITY AND OPERATION RT&A # () (Vol.) March RELIABILITY RIK AND AVAILABILITY ANLYI OF A CONTAINER GANTRY CRANE

More information

Hosted Voice Self Service Installation Guide

Hosted Voice Self Service Installation Guide Hosted Voce Self Servce Installaton Gude Contact us at 1-877-355-1501 learnmore@elnk.com www.earthlnk.com 2015 EarthLnk. Trademarks are property of ther respectve owners. All rghts reserved. 1071-07629

More information

A Secure Password-Authenticated Key Agreement Using Smart Cards

A Secure Password-Authenticated Key Agreement Using Smart Cards A Secure Password-Authentcated Key Agreement Usng Smart Cards Ka Chan 1, Wen-Chung Kuo 2 and Jn-Chou Cheng 3 1 Department of Computer and Informaton Scence, R.O.C. Mltary Academy, Kaohsung 83059, Tawan,

More information

A DATA MINING APPLICATION IN A STUDENT DATABASE

A DATA MINING APPLICATION IN A STUDENT DATABASE JOURNAL OF AERONAUTICS AND SPACE TECHNOLOGIES JULY 005 VOLUME NUMBER (53-57) A DATA MINING APPLICATION IN A STUDENT DATABASE Şenol Zafer ERDOĞAN Maltepe Ünversty Faculty of Engneerng Büyükbakkalköy-Istanbul

More information

NPAR TESTS. One-Sample Chi-Square Test. Cell Specification. Observed Frequencies 1O i 6. Expected Frequencies 1EXP i 6

NPAR TESTS. One-Sample Chi-Square Test. Cell Specification. Observed Frequencies 1O i 6. Expected Frequencies 1EXP i 6 PAR TESTS If a WEIGHT varable s specfed, t s used to replcate a case as many tmes as ndcated by the weght value rounded to the nearest nteger. If the workspace requrements are exceeded and samplng has

More information

Network Security Situation Evaluation Method for Distributed Denial of Service

Network Security Situation Evaluation Method for Distributed Denial of Service Network Securty Stuaton Evaluaton Method for Dstrbuted Denal of Servce Jn Q,2, Cu YMn,2, Huang MnHuan,2, Kuang XaoHu,2, TangHong,2 ) Scence and Technology on Informaton System Securty Laboratory, Bejng,

More information

Analysis of Premium Liabilities for Australian Lines of Business

Analysis of Premium Liabilities for Australian Lines of Business Summary of Analyss of Premum Labltes for Australan Lnes of Busness Emly Tao Honours Research Paper, The Unversty of Melbourne Emly Tao Acknowledgements I am grateful to the Australan Prudental Regulaton

More information

Traffic-light a stress test for life insurance provisions

Traffic-light a stress test for life insurance provisions MEMORANDUM Date 006-09-7 Authors Bengt von Bahr, Göran Ronge Traffc-lght a stress test for lfe nsurance provsons Fnansnspetonen P.O. Box 6750 SE-113 85 Stocholm [Sveavägen 167] Tel +46 8 787 80 00 Fax

More information

An Enhanced Super-Resolution System with Improved Image Registration, Automatic Image Selection, and Image Enhancement

An Enhanced Super-Resolution System with Improved Image Registration, Automatic Image Selection, and Image Enhancement An Enhanced Super-Resoluton System wth Improved Image Regstraton, Automatc Image Selecton, and Image Enhancement Yu-Chuan Kuo ( ), Chen-Yu Chen ( ), and Chou-Shann Fuh ( ) Department of Computer Scence

More information

To manage leave, meeting institutional requirements and treating individual staff members fairly and consistently.

To manage leave, meeting institutional requirements and treating individual staff members fairly and consistently. Corporate Polces & Procedures Human Resources - Document CPP216 Leave Management Frst Produced: Current Verson: Past Revsons: Revew Cycle: Apples From: 09/09/09 26/10/12 09/09/09 3 years Immedately Authorsaton:

More information

Project Networks With Mixed-Time Constraints

Project Networks With Mixed-Time Constraints Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa

More information

M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS

M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS Bogdan Cubotaru, Gabrel-Mro Muntean Performance Engneerng Laboratory, RINCE School of Electronc Engneerng Dubln Cty

More information

Single and multiple stage classifiers implementing logistic discrimination

Single and multiple stage classifiers implementing logistic discrimination Sngle and multple stage classfers mplementng logstc dscrmnaton Hélo Radke Bttencourt 1 Dens Alter de Olvera Moraes 2 Vctor Haertel 2 1 Pontfíca Unversdade Católca do Ro Grande do Sul - PUCRS Av. Ipranga,

More information

Efficient Project Portfolio as a tool for Enterprise Risk Management

Efficient Project Portfolio as a tool for Enterprise Risk Management Effcent Proect Portfolo as a tool for Enterprse Rsk Management Valentn O. Nkonov Ural State Techncal Unversty Growth Traectory Consultng Company January 5, 27 Effcent Proect Portfolo as a tool for Enterprse

More information

Learning the Best K-th Channel for QoS Provisioning in Cognitive Networks

Learning the Best K-th Channel for QoS Provisioning in Cognitive Networks 000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050

More information

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS Chrs Deeley* Last revsed: September 22, 200 * Chrs Deeley s a Senor Lecturer n the School of Accountng, Charles Sturt Unversty,

More information

Linear Circuits Analysis. Superposition, Thevenin /Norton Equivalent circuits

Linear Circuits Analysis. Superposition, Thevenin /Norton Equivalent circuits Lnear Crcuts Analyss. Superposton, Theenn /Norton Equalent crcuts So far we hae explored tmendependent (resste) elements that are also lnear. A tmendependent elements s one for whch we can plot an / cure.

More information

Performance Analysis and Coding Strategy of ECOC SVMs

Performance Analysis and Coding Strategy of ECOC SVMs Internatonal Journal of Grd and Dstrbuted Computng Vol.7, No. (04), pp.67-76 http://dx.do.org/0.457/jgdc.04.7..07 Performance Analyss and Codng Strategy of ECOC SVMs Zhgang Yan, and Yuanxuan Yang, School

More information

The Use of Analytics for Claim Fraud Detection Roosevelt C. Mosley, Jr., FCAS, MAAA Nick Kucera Pinnacle Actuarial Resources Inc.

The Use of Analytics for Claim Fraud Detection Roosevelt C. Mosley, Jr., FCAS, MAAA Nick Kucera Pinnacle Actuarial Resources Inc. Paper 1837-2014 The Use of Analytcs for Clam Fraud Detecton Roosevelt C. Mosley, Jr., FCAS, MAAA Nck Kucera Pnnacle Actuaral Resources Inc., Bloomngton, IL ABSTRACT As t has been wdely reported n the nsurance

More information

Course outline. Financial Time Series Analysis. Overview. Data analysis. Predictive signal. Trading strategy

Course outline. Financial Time Series Analysis. Overview. Data analysis. Predictive signal. Trading strategy Fnancal Tme Seres Analyss Patrck McSharry patrck@mcsharry.net www.mcsharry.net Trnty Term 2014 Mathematcal Insttute Unversty of Oxford Course outlne 1. Data analyss, probablty, correlatons, vsualsaton

More information

Methodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications

Methodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications Methodology to Determne Relatonshps between Performance Factors n Hadoop Cloud Computng Applcatons Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng and

More information

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,

More information

METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS

METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng

More information

Inter-Ing 2007. INTERDISCIPLINARITY IN ENGINEERING SCIENTIFIC INTERNATIONAL CONFERENCE, TG. MUREŞ ROMÂNIA, 15-16 November 2007.

Inter-Ing 2007. INTERDISCIPLINARITY IN ENGINEERING SCIENTIFIC INTERNATIONAL CONFERENCE, TG. MUREŞ ROMÂNIA, 15-16 November 2007. Inter-Ing 2007 INTERDISCIPLINARITY IN ENGINEERING SCIENTIFIC INTERNATIONAL CONFERENCE, TG. MUREŞ ROMÂNIA, 15-16 November 2007. UNCERTAINTY REGION SIMULATION FOR A SERIAL ROBOT STRUCTURE MARIUS SEBASTIAN

More information

Abstract. 260 Business Intelligence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING

Abstract. 260 Business Intelligence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING 260 Busness Intellgence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING Murphy Choy Mchelle L.F. Cheong School of Informaton Systems, Sngapore

More information

Quantization Effects in Digital Filters

Quantization Effects in Digital Filters Quantzaton Effects n Dgtal Flters Dstrbuton of Truncaton Errors In two's complement representaton an exact number would have nfntely many bts (n general). When we lmt the number of bts to some fnte value

More information

J. Parallel Distrib. Comput.

J. Parallel Distrib. Comput. J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n

More information

VoIP over Multiple IEEE 802.11 Wireless LANs

VoIP over Multiple IEEE 802.11 Wireless LANs SUBMITTED TO IEEE TRANSACTIONS ON MOBILE COMPUTING 1 VoIP over Multple IEEE 80.11 Wreless LANs An Chan, Graduate Student Member, IEEE, Soung Chang Lew, Senor Member, IEEE Abstract IEEE 80.11 WLAN has hgh

More information

1 Example 1: Axis-aligned rectangles

1 Example 1: Axis-aligned rectangles COS 511: Theoretcal Machne Learnng Lecturer: Rob Schapre Lecture # 6 Scrbe: Aaron Schld February 21, 2013 Last class, we dscussed an analogue for Occam s Razor for nfnte hypothess spaces that, n conjuncton

More information

An RFID Distance Bounding Protocol

An RFID Distance Bounding Protocol An RFID Dstance Boundng Protocol Gerhard P. Hancke and Markus G. Kuhn May 22, 2006 An RFID Dstance Boundng Protocol p. 1 Dstance boundng Verfer d Prover Places an upper bound on physcal dstance Does not

More information

SIMPLE LINEAR CORRELATION

SIMPLE LINEAR CORRELATION SIMPLE LINEAR CORRELATION Smple lnear correlaton s a measure of the degree to whch two varables vary together, or a measure of the ntensty of the assocaton between two varables. Correlaton often s abused.

More information

Rapid Estimation Method for Data Capacity and Spectrum Efficiency in Cellular Networks

Rapid Estimation Method for Data Capacity and Spectrum Efficiency in Cellular Networks Rapd Estmaton ethod for Data Capacty and Spectrum Effcency n Cellular Networs C.F. Ball, E. Humburg, K. Ivanov, R. üllner Semens AG, Communcatons oble Networs unch, Germany carsten.ball@semens.com Abstract

More information

Descriptive Models. Cluster Analysis. Example. General Applications of Clustering. Examples of Clustering Applications

Descriptive Models. Cluster Analysis. Example. General Applications of Clustering. Examples of Clustering Applications CMSC828G Prncples of Data Mnng Lecture #9 Today s Readng: HMS, chapter 9 Today s Lecture: Descrptve Modelng Clusterng Algorthms Descrptve Models model presents the man features of the data, a global summary

More information