An Optimal Test Pattern Selection Method to Improve the Defect Coverage
|
|
- Garry Byrd
- 5 years ago
- Views:
Transcription
1 A Optimal Test Patter Selectio Method to Improve the Defect Coverage Yuxi Tia, Michael R. Grimaila, Weipig Shi ad M. Ray Mercer Departmet of Electrical Egieerig Texas A&M Uiversity, College Statio, Texas 77843, USA Fax: {ytia, wshi, Departmet of Systems ad Egieerig Maagemet Air Force Istitute of Techology, Wright-Patterso AFB, Ohio 45433, USA Abstract It is well kow that -detectio test sets are effective to detect umodeled defects ad improve the defect coverage. However, i these sets, each of the -detectio test patters has the same importace o the overall test set performace. I other words, the test patter that detects a fault for the first time plays the same importat role as the test patter that detects that fault for the ()- th time. This is ot a accurate assumptio, especially as the deep submicro techologies are widely used today. I this paper, we propose a liear programmigbased optimal test patter selectio method which aims at reducig the overall defect part level (DPL). Usig resistive bridge faults as surrogates, our experimetal results o ISCAS85 circuits demostrate the proposed test patter selectio method achieves a higher defect coverage tha traditioal -detectio method. 1 Itroductio Semicoductor maufacturers strive to attai a high yield (ideally 100%) whe fabricatig itegrated circuits. Ufortuately, umerous factors ca lead to a variety of maufacturig defects which may reduce the overall yield. The purpose of testig is to idetify ad elimiate ay defective chips after the chips are maufactured. However, it is curretly impractical to test exhaustively for all possible defects. This is a result of the computatioal ifeasibility of accurately modelig defects, limitatios imposed by existig maufacturig test equipmet ad time/ecoomic costraits imposed by the test egieers. For these reasos, the stuck-at-fault (SAF ) model has bee accepted as the stadard model to geerate test patters. Most of the existig commercial ATPG tools use the SAF coverage as a metric of the quality of a test set ad termiate test geeratio whe a high SAF fault coverage is attaied. Although sigle stuck-at fault detectio is widely accepted i idustry, it ca ot detect may other types of defects [19]. Especially as the widespread usage of deep submicro techology, the probability of developmet of trasiet ad patter sesitive faults is risig tremedously. To address the deficiecy of stuck-at fault detectio, oe approach is to build more sophisticated fault models, such as the bridge fault model [15][12] ad the delay fault model [10]. However, sice most existig ATPG tools are desiged for sigle stuck-at faults, cosiderable effort must be made to exted these tools to the ew fault models. I this paper, we follow aother approach: -detectio, where each sigle stuck-at fault is detected multiple times i order to catch other types of defects [18] [8]. I the -detectio approach, the existig ATPG tools ca be used with miimum efforts to achieve the best results. The first work o -detectio was doe by Ma, Fraco ad McCluskey [14]. I a experimet, they showed that as the umber of uique detectios for each fault icreases, the defect coverage always improves compared with other test geeratio schemes. Whe is sufficietly large, the -detectio stuck-at test patter sets lead to the idetificatio of almost all defective chips. Subsequet works [16][20] have also provided the aalysis of the effectiveess of -detectio test patter sets. However, the set of test patters that detect each stuck-at fault times is ofte too large, resultig i both high tester memory usage ad log testig time. This is because the -detectio method treats all test patters that detect a stuck-at fault equally importat whe the fault is detected less tha times. Previous research has proved that ot all the patters i the test sets are equally efficiet i reducig the defect coverage [9]. Thus, -
2 detectio o each sigle stuck-at fault would somehow waste some test patters ad reduce the performace of the test set. Therefore, a post test geeratio compactio or optimizatio is ecessary. I this paper, we propose a liear programmigbased ew test patter geeratio strategy which selects the optimal subset of test patters from the -detectio test set based o a weighted defect part level estimatio model MPG-D model [4]. Istead of geeratig the test patters directly, our method aims at selectig a optimal patter set from a big -detectio set. We use weighted MPG-D model as our objective fuctio to select the patters, which accurately reflect the cotributios of each patter to detect umodeled defects. Sice the accurate defect coverage is almost impossible to get [1], we use the surrogate detectio approach to compare the fault coverage of the test patter set with the -detectio test patter set. (Here, surrogates mea the fault models that are ot targeted directly durig the test patter geeratio but are used to evaluate the quality of the test patter sets). Uder this approach, the test patters are geerated for a specific target fault model (sigle stuck-at fault) ad the simulated o the surrogate fault model (bridge fault). The rest of this paper is orgaized as follows: I sectio 2, we itroduce the weighted MPG-D defect part level estimatio model. Sectio 3 itroduces the applicatio of liear programmig to select optimal test patters from a -detectio test set. Sectio 4 summarizes the experimetal results ad compares the surrogate fault coverage of our method ad that of the - detectio method o ISCAS85 bechmark circuits. Fially, i sectio 5, we coclude our research results. 2 Defect Part Level Estimatio Model To geerate a test set, test geerators always use some objective fuctios to determie if the fial objective is achieved. For the sigle stuck-at fault orieted ATPG tools, the objective fuctio is the sigle stuck-at fault coverage, ad the objective is to maximize, to 100% if possible, the fault coverage. For the -detectio methods, the objective is to detect each stuck-at fault by differet patters at least times. As log as -detectio o a fault is achieved, this fault is dropped from the fault list ad o loger targeted uder additioal tests. This process ca be explaied by the followig model: Give a set of test patters T = {t 1, t 2,..., t } ad fault set F = {f 1, f 2,..., f m }, defie the objective fuctio as: M(F, T ) = m M(f i, T ) where 1 if f i is detected less tha times by T M(f i, T ) = 0 otherwise. The objective of the -detectio methods is to geerate a set T so that the objective fuctio M(F, T ) is miimized. However, this step objective fuctio of -detectio is ot optimal because as the umber of detectios icreases, the objective fuctio should gradually decrease sice as more defects are detected at each site, the probability of excitig oe of the remaiig defects becomes smaller. Cosider Figure 1, the etire test space available at a test site i the circuit is represeted as a box i the Ve diagram. Each oval represets the portio of that test space which will excite a particular defect at that test site (give that the site is observed). Thus, a defect is detected whe a test is chose so that it falls withi that defect s correspodig oval (Multiple defects may i some cases be detected by the same test patter). Oce a defect is detected, detectig it agai will ot further reduce the defect part level, therefore, defect s oval o loger appears amog those udetected defects. At ay particular poit i time, the probability of excitig a as yet udetected defect give that the site is observed is simply the ratio of the area of the test space covered by the ovals to the total area. Thus, the probability of excitig a as yet udetected defect teds to decrease as test patters are applied ad defects are detected. Based o this observatio, a defect part level estimatio model, MPG-D model, was proposed [3]. Ulike other defect part level estimatio models which deped upo fault coverage, it does ot go to zero whe fault coverage reaches 100%. I additio, it represets the fact that the probability of excitatio decreases as the umber of site observatios icreases ad that more tha oe defect may be preset at ay fault site. The origial MPG-D model defies the relatioship betwee the umber of times a site is observed ad its cotributio to the DPL. First of all, this model assumes the uiform distributio of all the defects across the fault sites. Thus, for each fault site j, a iitial cotributio to the overall DPL is defied as: DP L j (0) = 1 Y No. of Fault Sites (1) where Y is the maufacturig yield. After the applicatio of a test patter sequece T = {p 1, p 2,..., p k }, where k 1, some fault sites have bee observed, some have bee excited, ad some defects have bee detected. Thus, refer to Figure 1, the probability of excitig a udetected defect at a site give that site is observed would
3 ## ### ### ### ### ### ### ### ### ### ### ### ### ### ### # $$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $$$ $ % & %% && %% && %% && %% && '''''''' '''''''' '''''''' '''''' (((((((( (((((((( (((((((( (((((( ''''' ((((( ))))) ))))) )))))) )))))) )))))) ))))) ))))) ))) %% ***** ***** ****** ****** ****** ****** ***** *** && ))) ***%% && %% && %% && # $!!!!!!! """"""" %% &&!!!!!! """""" % & Iitial Test Space Reduced Test Space After Test Patter 1 Applied Reduced Test Space After 2 Test Patters Applied Test Patter 1 Test Patter 2 Figure 1. The probability of excitatio decreases as testig progresses be reduced as more test patters are applied. This probability has bee studied ad show to be a decayig expoetial fuctio of the umber of times that site has bee observed previously ad a time costat φ [3]: P excite = e obs j φ (2) where obs j meas the total umber of observatios of test site j. Sice the idividual probabilities are disjoit, the overall DPL is merely the sum of the cotributio from each idividual site. Therefore, DP L(T ) = No. of Fault Sites DP L j (k) (3) Ad for this test set T with k patters (k 1), each site s defect part level cotributio is give i the followig equatio [3]: DP L j (k) = DP L j (k 1)(1 A e obs j φ ) obs jk (4) where A ad φ are two costats related to the maufacturer s process ad circuits, ad DP L j (k 1) is the Number of Fault Sites Histogram of Couplig Capacitace for Circuit C432 Class III Class II Class I Couplig Capacitace (ff) Figure 2. Weight Fuctio Geeratio defect part level cotributio whe the first k 1 test patters are applied to the circuit uder test. The value obs j meas the total umber of observatios of test site j so far. The value of obs jk is 1 if fault j is observed by vector k, otherwise it is 0. Although the origial MPG-D model ca be used to predict the DPL, it suffers from the limitatio assumptio that all sites i the circuit are equally sesitive to defects. I reality, we fid that each site i the circuit has differet physical ad geometric characteristics that make it more or less susceptible to a uiform distributio of defects. Therefore, each site s cotributio to the overall DPL is ot uiform but istead is a strog fuctio of its physical layout. Larger sites with more die area are more likely to suffer from defects tha smaller sites. I order to create a more accurate model, we must accout for the physical characteristics as represeted i the layout of the circuit. I this paper, we will use a refied MPG-D model to estimate the DPL resultig from the applicatio of a give test set. We refer to this model as the weighted MPG-D sice we will assig differet weight for each site i the circuit based upo the physical ad/or electrochemical characteristics of each site. I geeral, the weights may be a fuctio of all the characteristic such as (but ot limited to) the physical area of the site, proximity to adjacet sites, proximity to active areas, io implatatio eergy, diffusio gradiet, etc. As recet studies have show [2][13], i the deep submicro techology era, the problems due to icreasig couplig capacitace have a sigificat adverse effect o the proper fuctio ad performace of VLSI system. Therefore, for simplicity, i this paper we will employ the couplig capacitace extracted from the layout as a idicator to geerate the weight fuctio ad refie the origial MPG-D model. First, the circuit layout was geerated with Cadece Silico Esemble i TSMC 250m 3V 3-metal techology. Commercial par-
4 asitic extractio tools are used to extract the couplig capacitace of each site. To simplify the problem without losig geerality, we classify the fault sites ito several classes based o their greatest couplig capacitace, for the classes with bigger couplig capacitace, we assig a bigger weight, for the classes with smaller couplig capacitace, we assig a smaller wight. To show these steps i detail, the histogram of couplig capacitace for circuit C432 fault sites is show i Figure 2. For this circuit, we set up two thresholds for to separate the fault sites ito three classes, if the couplig capacitace of a fault site is bigger tha the first threshold (1.5fF), we put them ito class I; if the couplig capacitace is betwee the secod threshold (1fF) ad the first threshold (1.5fF), we put them ito class II; For all other sites, we put them ito class III. Therefore the weight fuctio is defied i the followig step fuctio: w 1 if fault site j is i class I w j = w 2 if fault site j is i class II otherwise. w 3 where w 1 w 2 w 3. I our experimets, we choose the weights betwee 1.0 ad Now we have the weighted MPG-D model defied i equatio (5): DP L(T ) = No. of Fault Sites w j DP L j (k) (5) The solid lie i Figure 3 shows a example of DPL cotributio estimated by equatio (4) with the weight of 1.0. The curve shows that as the umber of observatios o test site j icreases, the cotributio to the overall defect part level decreases as a expoetial fuctio. Now we have derived the DPL estimatio fuctio show i equatio (5), our objective is to geerate a test patter set so that the DPL is miimized. Istead of geeratig the test patters directly, our method selects a optimal patter set from a big -detectio superset. The -detectio superset are geerated by ruig the stadard ATPG tools repeatedly util every fault is detected times. As we discussed before, although -detectio test set could effectively reduce the defect part level, the umber of geerated patters is much larger tha covetioal methods. Limitatios o tester memory size ad the high cost of testig time severely limits the umber of test patters that ca be applied to a device i a commercial settig. As a result, a post processig optimizatio process is required to select the best subset of patters from the -detectio superset to use as the actual test set. I the ext sectio, we are goig to itroduce a liear programmig based method to select the patters from this big -detectio test patter set. DPL Cotributio from a Fault Site (ppm) A B DPL Cotributio Liear Approximatio No. of Observatios o A Idividual Fault Site Figure 3. Liearizig the DPL Reductio Curve 3 Optimal Test Patter Selectio I this sectio, we develop a liear programmig method to solve the optimal test patter selectio problem discussed i sectio 2. Liear programmig is beig widely used today i a wide variety of optimizatio problems, icludig the VLSI desig ad testig problems [5][7]. The fial goal of our liear programmig model is to select a optimal or ear-optimal set of test patters from a superset with the objective that the objective fuctio (5) is miimized. Sice the DPL fuctio defied i (5) is ot a liear oe, it is ot easy to miimize this objective fuctio directly. Therefore, we first liearize the weighted MPG- D model i terms of equatios ad costraits. Cosider the DPL cotributio curve show i Figure 3 which shows the relatioship betwee the umber of detectios for each fault ad the expected value of the objective fuctio for a specific fault, if we wat to geerate the optimal test patter set which results i the miimum expected value, we ca liearize the weighted MPG-D curve with the three liear segmets show i Figure 3 to approximate the real objective fuctio, ad this approach will sigificatly improve the solvability of the origial problem without sacrificig much accuracy. From the previous discussio, we ca defie the problem as follows: Give a fault dictioary Π T F, where T = {t 1, t 2,..., t } is the big -detectio set geerated by ruig ATPG tools repeatedly ad F = {f 1, f 2,..., f m } is the fault set. If (t i, f j ) Π, patter t i detects fault f j. The problem is how to select a set T T so that T u, where u is the upper boud allowed for set T, ad at the same time, the defect part level is miimized. Therefore, the objective fuctio of C
5 the problem is defied as: Mi : DP L(T ) = m w j DP L j (v) (6) where m is the umber of stuck-at faults i the circuit, w j is the weight assiged to fault j, v is the umber of detectios o fault f j ad DP L j (v) is the idividual defect part level cotributio from fault site j. Sice DP L(T ) is ot a liear fuctio, it is ot so easy to fid the optimal solutio of objective fuctio (6) directly. Therefore, we ca approximate DP L j (v) to a liear fuctio DP L j (v) so that we ca use the stadard liear programmig method to solve it. The the objective fuctio becomes: Mi : DP L(T ) = m w j DP L j (v) (7) where DP L j (v) icludes the three liearized segmets show i Figure 3. These three lies correspod to three liear fuctios (here, they are amed as M 1, M 2 ad M 3, respectively) as follows: M 1 = DP L j (v) = M 2 = DP Lj(b) DP Lj(0) b v + DP L j (0) (If 0 v < b) DP Lj(c) DP Lj(b) c b (v b) +DP L j (b) (If b v < c) M 3 = 0 (If v c) where v is the umber of detectios o a fault, b ad c are the x-coordiate values of the poits B ad C i Figure 3. We ca also fid that i each segmet, (e.g., 0 v < b, b v < c ad v c), DP L j (v) is always the maximum value of the three fuctios defied i (8). Therefore, for all v 0, the three equatios i (8) are equivalet to the followig objective fuctio: DP L j (v) = Maximum{M 1, M 2, M 3 } (9) I the fault dictioary, for ay t i T, we defie a idicator variable x i such that if t i is i T, x i = 1; otherwise, x i = 0. Therefore, { 1 If ti T x i = (10) 0 If t i / T Sice the total umber of patters i T is limited to u, we have x i u (11) (8) By usig the idicator variable, we ca easily fid that v, which is the umber of detectios o a fault ca be expressed as: v = x i (12) (t i,f j) Π Here, i equatios from (6) to (12), 1 i ad 1 j m. Combiig equatios (7), (8), (9), (10), (11) ad (12) together, we ca trasform the origial optimizatio problem ito a iteger liear programmig problem sice x i must be a iteger. This problem ca be solved by iteger liear programmig algorithm. However, iteger liear programmig problem is kow a NP-complete problem without ay efficiet solutios. There are several ways to trasfer them ito a o-iteger problem, such as iterior poit method, differetiable method ad Taylor approximatio method [17][21]. I this paper, we use the relaxatio ad roudig method to obtai a o-iteger liear programmig model. First of all, the iteger idicators x 1, x 2,..., x i the origial problem are relaxed ad redefied. After the problem is solved, we will roud these variables back to itegers. So, we first covert the costrait (10) ad chage it to a problem that ca be solved more easily. We redefie x i as the probability of test patter t i to be selected from superset T. Thus, for every 1 i, we form the followig costrait: 0 x i 1,where 1 i (13) With this trasform, we ca fialize our optimal patter selectio problem ito the followig liear programmig model: Mi: DP L(T ) = Subject to: m w j DP L j (v) DP L j (v) DP L j(b) DP L j (0) b +DP L j (0) DP L j (v) DP L j(c) DP L j (b) c b DP L j (v) 0 x i u +DP L j (b) [ (t i,f j) Π (t i,f j) Π x i x i b] 0 x i 1 (14)
6 where 1 i, 1 j m. Compared with the origial iteger liear programmig oe, this liear programmig problem is much easier to solve. After this problem is solved, we eed to roud the solutio of x i back to 0 or 1 sice we assume it is the probability of selectig test patter t i from the superset T. Sice the defiitio of x i is chaged, we eed to verify that the fial result of the x i satisfies the origial coditios (10) ad (11). Suppose the solutios of the liear programmig model above are sets {ˆx i 1 i } ad { DP L j (v) 1 j m}. The based o the defiitio of x i, we ca roud x i back to 0 or 1 by followig the distributios: Probability[x i = 1] = ˆx i Probability[x i = 0] = 1 ˆx i (15) This distributio ca be accomplished i the followig way: 1. Geerate a radom umber r; 2. Compare r ad ˆx i ; 3. If r ˆx i, the x i is set to 1, otherwise, x i is set to 0. By roudig the result set ˆx i i this way, the origial costrait equatio (10) is satisfied. Now, we eed to cosider the origial costrait (11). It is possible that this costrait is violated; However, based o the probability theory [6], if we cosider the expected value of the results, we ca get the followig equatio: E[ x i ] = = = E[x i ] P [x i = 1] ˆx i u (16) I the derivatio here, the first equality is because the operator is liear, eve for radom variables that are depedet, ad the third oe is from equatio (15). Therefore, from equatio (16), we ca see that costrait (11) is satisfied o the average. That meas the expected value of the result is limited to the upper boud of the set T, which satisfies the costrait (11) of the origial problem. From the discussio i this sectio, we have set up a liear programmig model which attempts to obtai the miimum objective fuctio defied i sectio 2 for a test patter set selected from a superset. The whole optimizatio method ca be summarized as the follows: First, a -detectio test patter set as the test patter superset T ad the correspodig fault dictioary Π = T F are geerated, the the liear programmig model which is discussed i this sectio is formed ad solved. After the liear programmig problem is solved, the result is rouded to select the set T from the origial superset T. Fially, the resistive bridge fault is used as the surrogates ad the set selected is iput to the surrogate fault simulator PROBE [11] to get the resistive bridge fault coverage as the idicator of the test quality. 4 Experimetal Results The liear programmig problem preseted i this paper were solved usig a AMD Athlo XP 2800 PC ruig the Redhat Liux operatig system; The bridge fault simulator PROBE are workig o a SUN Ultra-4 workstatio ruig Solaris 5.7 operatig system. Stadard -detectio test patters were geerated by ruig Metor FastSca T M, a commercial ATPG tool, repeatedly. The liear programmig problem is solved by So- Plex 1.2.1, a liear programmig solver based o sequetial object-orieted simplex algorithm writte i the C laguage [22]. 4.1 Experimet Desig We recogize that it is virtually impossible to get the actual defect coverage of a test set without itetioally iducig defects i each device ad coductig a detailed decostructio of all devices. However, to show the effectiveess of liear programmig based patter selectio methodology, we desiged a experimet that would compare the surrogate fault coverage betwee - detectio method ad patter selectio method we proposed. I the experimet, there are two idepedet variables (circuit ad umber of patters geerated) ad the resultig surrogate fault coverage is compared. Our experimets use the te o-trivial ISCAS85 bechmark circuits. For each circuit, we geerate a collapsed stuckat fault list ad ru stadard ATPG tool repeatedly to get a -detectio superset (here i the superset, is set to be 15). Meawhile, 2-detectio set to 7-detectio test sets are also geerated by the same method. The, we use the patter selectio methodology discussed before to select the test patters from this 15-detectio superset ad geerate the set of test patters, whose sizes are limited to the 2-detectio to 7-detectio test sets, respectively. Whe all the test sets are ready, we use PROBE simulator to do the resistive bridge fault simulatio ad get the fault coverage as the fial results. We are also iterested i varyig the umber of test patters to reveal the trade-offs betwee test set size ad the resultig surrogate fault coverage. Here, we geerate a test patter set which has 100% sigle stuck-at fault coverage ad use the umber of test patters geerated as the base umber k. Specifically, the test set size limits are varied from oe times (k) to ie times (9k)
7 Detectio Detectio Figure 4. Bridge fault coverage vs. test set size for circuit c432 Figure 6. Bridge fault coverage vs. test set size for circuit c Detectio Detectio Figure 5. Bridge fault coverage vs. test set size for circuit c2670 Figure 7. Bridge fault coverage vs. test set size for circuit c5315 the umber of test patters that would be applied usig stadard best practice ATPG tool. Subsequetly, we use the ATPG tool to geerate a 15-detectio test patter superset. From this superset, subsets of various sizes were selected for each circuit. The the bridge fault simulatio is performed to get the fial bridge fault coverage. For each of the circuits, test geeratio methodologies, ad umber of patters geerated, we used PROBE resistive bridge fault simulator to get the bridge fault coverage as the idicator of the defect coverage Detectio Experimet Results Figures 4 to 8 shows the resistive fault coverage of the differet test patter size o differet circuits ad methods. is the umber we used i -detectio method, i these figures, chages from 2 to 7, which meas each Figure 8. Bridge fault coverage vs. test set size for circuit c7552
8 Table 1. CPU time to geerate the 15-detectio superset ad to select optimal test patters Circuit CP U 15 det CP U opt T overhead c s 1.9s 1.7% c s 3.2s 2.6% c s 9.5s 4.2% c s 18.6s 4.1% c s 132.7s 12.3% c s 198.5s 14.1% c s 320.4s 16.2% c s 579.3s 28.8% c s 386.9s 25.9% c s s 31.7% fault is detected 2 to 7 times. The umbers o the top of the colums idicate the resistive fault coverage for two differet methods. Based o the superset test patters umber of patters geerated by -detectio method, a optimal test patter set is selected such that it has the same umber of patters as -detectio. The other circuits have the similar results as these figures show. The results show here verify that optimal patter selectio method always get a higher bridge fault coverage tha the -detectio methodology as the size of the test set is same. These results also cofirm the correctess of weighted MPG-D model. Whe the umber of patters is smaller, the surrogate fault coverage differece betwee the two methods is bigger, while with the umber of patters icrease, the differece is smaller. This is because with the icrease of the umber of patters i the test set, the improvemet o the resistive fault coverage is becomig harder ad harder to get, also the overlap betwee the two test sets is bigger sice our superset is a 15-detectio set geerated by the same -detectio methodology, thus, the differece is smaller for bigger i these figures. Table 1 shows the ruig time (i secods) of the proposed algorithm. Colum CP U 15 det idicates the CPU time used to geerate the 15-detectio superset, CP U opt shows the average CPU time used to select the optimal test patter sets ad colum T overhead shows the CPU time overhead. We ca observe from the table that depedig o the circuit size, the CPU time overhead is varyig from 2% to 31.7%. Test egieers are ofte faced with the critical questio: How may test patters should be applied to the device uder test?. Their decisio has a eormous impact o the quality ad profitability of the product. Idustry testig regimes typically cosist of a combiatio of DC parametric, IDDQ, fuctioal, ad sca stuck-at fault based tests. Applyig too may test patters will Optimal Selectio Method Fault Coverage -Detectio Fault Coverage No. of Patters Applied Figure 9. Resistive bridge fault coverage vs. test set size for circuit c880 slow dow the test time ad icrease the tester memory. But if the test patter set is too small, it is hard to detect the umodeled defects. I order to make a ratioal decisio, oe must take ito accout the cost of testig time ad the desired defect coverage, as well as the size of available tester memory. I Figure 9, we use circuit c880 as a example to show a graphical represetatio of the trade-offs i test quality i terms of surrogate fault coverage ad the test set size. For circuit c880, 1-detectio test set icludes 40 patters, we vary the umber of patters applied from k=40 to 9k=360. Here, both optimal patter selectio method ad -detectio method are used ad the results are compared, it is obvious that optimal patter selectio method always results i a higher fault coverage tha -detectio. Graphs such as this provide a visual guide to aid the test egieer i makig trade-off decisios. Similar experimets are also carried o the other ISCAS85 circuits. From this figure, we ca get the umber of patters geerated by two differet methods whe the same bridge fault coverage is achieved. 5 Coclusios I this paper, we preseted a optimal test patter selectio methodology based o the weighted MPG-D model. The couplig capacitace value extracted from the bechmark circuit layout is used as the idicator to select the weight for each fault site to estimate the DPL. The objective of the methodology is to select a subset of test patters from a superset that results i the optimal defect coverage for a give, fixed test size. Compariso of the method to -detectio showig the effectiveess of optimal patter selectio at icreasig defect coverage. The primary result of this research is the icrease of the resistive bridge fault coverage. We achieved this result by the formulatio of a liear programmig model
9 which was the solved based o appropriate costraits for the give circuit. The experimetal results proved that this method is effective i detectig the umodeled defects ad thus improve the defect coverage. We believe that the icrease i the surrogate fault coverage will allow the methodology to be icorporated ito commercial practice. Fially, a graphical represetatio of the trade-off betwee the test quality ad test size is preseted to aid the test egieer i makig decisios regardig trade-offs betwee test set quality ad test set size. 6 Ackowledgmet The authors thak Dr. D. M. H. Walker for providig bridge fault simulator PROBE. Refereces [1] M. L. Bushell ad V. D. Agrawal Essetials of Electroic Testig for Digital, Memory ad Mixed-Sigal VLSI Circuits, Klumer Academic Publishers, [2] Kwag-Tig Cheg, S. Dey, M. Rodgers, K. Roy, Test challeges for deep sub-micro techologies, Proc. Desig Automatio Coferece, pp , [3] J. Dworak, M. R. Grimaila, S. Lee, L. C. Wag ad M. R. Mercer, Ehaced DO-RE-ME Based Defect Level Predictio Usig Defect Site Aggregatio - MPG-D, Proc. Iteratioal Test Coferece, pp , [4] J.Dworak, J.D.Wicker, S.Lee, M.R.Grimaila, K.M.Bulter, B.Stewart, L.C.Wag ad M.R.Mercer. Defect-Orieted Testig ad Defective Part Level Predictio. IEEE Desig ad Test of Computers, 31-41, Ja.-Feb [5] F. Fallah, S. Devadas ad K. Keutzer, Fuctioal vector geeratio for HDL models usig liear programmig ad boolea satisfiability, IEEE Tras. o CAD of Itegrated Circuits ad Systems, 20(8), pp , Aug [6] W. Feller, A Itroductio to Probability Theory ad Its Applicatios, 3rd Editio, Joh Wiley ad Sos, New York, [7] P. F. Flores, H. C. Neto ad J. P. Marques-Silva, A exact solutio to the miimum size test patter problem, ACM Tras. o Desig Automatio of Electroic Systems, 16(4), pp , Oct [8] M. R. Grimaila, S. Lee, J. Dworak, K. M. Bulter, B. Stewart, H. Balachadra, B. Houchis, V. Mathur, J. Park, L. C. Wag ad M. R. Mercer, RE-DO - radom excitatio ad determiistic observatio - first commercial experimet, Proc. VLSI Test Symposium, pp , [9] R. Kapur, J. Park ad M. R. Mercer, All tests for A fault are ot equally valuable for defect detectio, Proc. Iteratioal Test Coferece, pp , [10] A. Krstic ad K.-T. Cheg, Delay Fault Testig for VLSI Circuits, Kluwer Academic Publishers, [11] C. Y. Lee ad D. M. H. Walker, PROBE: A PPSFP simulator for resistive bridgig faults Proc. VLSI Test Symposium, pp , [12] Z. Li, X. Lu, W. Qiu, W. Shi ad H. Walker, A circuit level fault model for resistive opes ad bridges, ACM Tras. o Desig Automatio of Electroic Systems, Vol. 8, No. 4, pp , [13] Jig-Jia Liou, A. Krstic, Yi-Mig Jiag, Kwag-Tig Cheg Modelig, testig, ad aalysis for delay defects ad oise effects i deep submicro devices IEEE Tras. o CAD of Itegrated Circuits ad Systems, Vol. 22, No. 6, pp , [14] S. C. Ma, P. Fraco, ad E. J. McCluskey, A experimetal chip to evaluate test techiques experimet results, Proc. Iteratioal Test Coferece, pp , [15] S. D. Millma ad J. P. Garvery, Sr., A accurate bridgig fault test patter geerator, Proc. Iteratioal Test Coferece, pp , [16] I. Polia, I. Pomeraz, ad B. Becker, Exact computatio of maximally domiatig faults ad its applicatio to -detectio tests, Proc. Asia Test Symposium, pp.9-14, [17] P. Raghava, Probabilistic costructio of determiistic algorithms: approximatig packig iteger programs, J. of Computer ad System Sciece, Vol. 37, pp , [18] S. M. Reddy, I. Pomeraz ad S. Kajihara, O the effects of test compactio o defect coverage, Proc. VLSI Test Symposium, pp , [19] T. M. Storey, W.Maly, J.Adrews, ad M. Miske, Stuck fault ad curret testig compariso usig CMOS chip test, Proc. Iteratioal Test Coferece, pp , [20] C. Tseg, S. Mitra, S. Davidso ad E. J. McCluskey, A evaluatio of pseudo radom testig for detectig real defects, Proc. VLSI Test Symposium, pp , [21] R. J. Vaderbei, Liear Programmig: Foudatios ad Extesios, Kluwer Academic Publishers, Ju [22] R. Wuderlig, Paralleler ud Objektorietierter Simplex-Algorithmus, ZIB techical report TR 96-09, Berli, Germay, 1996.
A SELECTIVE POINTER FORWARDING STRATEGY FOR LOCATION TRACKING IN PERSONAL COMMUNICATION SYSTEMS
A SELETIVE POINTE FOWADING STATEGY FO LOATION TAKING IN PESONAL OUNIATION SYSTES Seo G. hag ad hae Y. Lee Departmet of Idustrial Egieerig, KAIST 373-, Kusug-Dog, Taejo, Korea, 305-70 cylee@heuristic.kaist.ac.kr
More informationApplication of Improved Genetic Algorithm to Two-side Assembly Line Balancing
206 3 rd Iteratioal Coferece o Mechaical, Idustrial, ad Maufacturig Egieerig (MIME 206) ISBN: 978--60595-33-7 Applicatio of Improved Geetic Algorithm to Two-side Assembly Lie Balacig Ximi Zhag, Qia Wag,
More informationCHAPTER 5 A NEAR-LOSSLESS RUN-LENGTH CODER
95 CHAPTER 5 A NEAR-LOSSLESS RUN-LENGTH CODER 5.1 GENERAL Ru-legth codig is a lossless image compressio techique, which produces modest compressio ratios. Oe way of icreasig the compressio ratio of a ru-legth
More informationAPPLICATION NOTE UNDERSTANDING EFFECTIVE BITS
APPLICATION NOTE AN95091 INTRODUCTION UNDERSTANDING EFFECTIVE BITS Toy Girard, Sigatec, Desig ad Applicatios Egieer Oe criteria ofte used to evaluate a Aalog to Digital Coverter (ADC) or data acquisitio
More informationRadar emitter recognition method based on AdaBoost and decision tree Tang Xiaojing1, a, Chen Weigao1 and Zhu Weigang1 1
Advaces i Egieerig Research, volume 8 d Iteratioal Coferece o Automatio, Mechaical Cotrol ad Computatioal Egieerig (AMCCE 7) Radar emitter recogitio method based o AdaBoost ad decisio tree Tag Xiaojig,
More informationA New Space-Repetition Code Based on One Bit Feedback Compared to Alamouti Space-Time Code
Proceedigs of the 4th WSEAS It. Coferece o Electromagetics, Wireless ad Optical Commuicatios, Veice, Italy, November 0-, 006 107 A New Space-Repetitio Code Based o Oe Bit Feedback Compared to Alamouti
More informationLogarithms APPENDIX IV. 265 Appendix
APPENDIX IV Logarithms Sometimes, a umerical expressio may ivolve multiplicatio, divisio or ratioal powers of large umbers. For such calculatios, logarithms are very useful. They help us i makig difficult
More informationx y z HD(x, y) + HD(y, z) HD(x, z)
Massachusetts Istitute of Techology Departmet of Electrical Egieerig ad Computer Sciece 6.02 Solutios to Chapter 5 Updated: February 16, 2012 Please sed iformatio about errors or omissios to hari; questios
More informationThe Detection of Abrupt Changes in Fatigue Data by Using Cumulative Sum (CUSUM) Method
Proceedigs of the th WSEAS Iteratioal Coferece o APPLIED ad THEORETICAL MECHANICS (MECHANICS '8) The Detectio of Abrupt Chages i Fatigue Data by Usig Cumulative Sum (CUSUM) Method Z. M. NOPIAH, M.N.BAHARIN,
More informationHybrid BIST Optimization for Core-based Systems with Test Pattern Broadcasting
Hybrid BIST Optimizatio for Core-based Systems with Test Patter Broadcastig Raimud Ubar, Masim Jeihhi Departmet of Computer Egieerig Talli Techical Uiversity, Estoia {raiub, masim}@pld.ttu.ee Gert Jerva,
More informationTest Time Minimization for Hybrid BIST with Test Pattern Broadcasting
Test Time Miimizatio for Hybrid BIST with Test Patter Broadcastig Raimud Ubar, Maksim Jeihhi Departmet of Computer Egieerig Talli Techical Uiversity EE-126 18 Talli, Estoia {raiub, maksim}@pld.ttu.ee Gert
More informationPROJECT #2 GENERIC ROBOT SIMULATOR
Uiversity of Missouri-Columbia Departmet of Electrical ad Computer Egieerig ECE 7330 Itroductio to Mechatroics ad Robotic Visio Fall, 2010 PROJECT #2 GENERIC ROBOT SIMULATOR Luis Alberto Rivera Estrada
More informationAC : USING ELLIPTIC INTEGRALS AND FUNCTIONS TO STUDY LARGE-AMPLITUDE OSCILLATIONS OF A PENDULUM
AC 007-7: USING ELLIPTIC INTEGRALS AND FUNCTIONS TO STUDY LARGE-AMPLITUDE OSCILLATIONS OF A PENDULUM Josue Njock-Libii, Idiaa Uiversity-Purdue Uiversity-Fort Waye Josué Njock Libii is Associate Professor
More informationA New Basic Unit for Cascaded Multilevel Inverters with the Capability of Reducing the Number of Switches
Joural of Power Electroics, ol, o, pp 67-677, July 67 JPE --6 http://dxdoiorg/6/jpe67 I(Prit: 98-9 / I(Olie: 9-78 A ew Basic Uit for Cascaded Multi Iverters with the Capability of Reducig the umber of
More informationA study on the efficient compression algorithm of the voice/data integrated multiplexer
A study o the efficiet compressio algorithm of the voice/data itegrated multiplexer Gyou-Yo CHO' ad Dog-Ho CHO' * Dept. of Computer Egieerig. KyiigHee Uiv. Kiheugup Yogiku Kyuggido, KOREA 449-71 PHONE
More informationSurvey of Low Power Techniques for ROMs
Survey of Low Power Techiques for ROMs Edwi de Agel Crystal Semicoductor Corporatio P.O Box 17847 Austi, TX 78744 Earl E. Swartzlader, Jr. Departmet of Electrical ad Computer Egieerig Uiversity of Texas
More informationData Acquisition System for Electric Vehicle s Driving Motor Test Bench Based on VC++ *
Available olie at www.sciecedirect.com Physics Procedia 33 (0 ) 75 73 0 Iteratioal Coferece o Medical Physics ad Biomedical Egieerig Data Acquisitio System for Electric Vehicle s Drivig Motor Test Bech
More informationFingerprint Classification Based on Directional Image Constructed Using Wavelet Transform Domains
7 Figerprit Classificatio Based o Directioal Image Costructed Usig Wavelet Trasform Domais Musa Mohd Mokji, Syed Abd. Rahma Syed Abu Bakar, Zuwairie Ibrahim 3 Departmet of Microelectroic ad Computer Egieerig
More informationPermutation Enumeration
RMT 2012 Power Roud Rubric February 18, 2012 Permutatio Eumeratio 1 (a List all permutatios of {1, 2, 3} (b Give a expressio for the umber of permutatios of {1, 2, 3,, } i terms of Compute the umber for
More informationX-Bar and S-Squared Charts
STATGRAPHICS Rev. 7/4/009 X-Bar ad S-Squared Charts Summary The X-Bar ad S-Squared Charts procedure creates cotrol charts for a sigle umeric variable where the data have bee collected i subgroups. It creates
More informationSuper J-MOS Low Power Loss Superjunction MOSFETs
Low Power Loss Superjuctio MOSFETs Takahiro Tamura Mutsumi Sawada Takayuki Shimato ABSTRACT Fuji Electric has developed superjuctio MOSFETs with a optimized surface desig that delivers lower switchig.
More informationConcurrent Fault Detection in Random Combinational Logic
Cocurret Fault Detectio i Radom Combiatioal Logic Petros Drieas ad Yiorgos Makris Departmets of Computer Sciece ad Electrical Egieerig Yale Uiversity Abstract We discuss a o-itrusive methodology for cocurret
More informationAnalysis and Optimization Design of Snubber Cricuit for Isolated DC-DC Converters in DC Power Grid
Aalysis ad Optimizatio Desig of Subber Cricuit for Isolated DC-DC Coverters i DC Power Grid Koji Orikawa Nagaoka Uiversity of Techology Nagaoka, Japa orikawa@st.agaokaut.ac.jp Ju-ichi Itoh Nagaoka Uiversity
More informationWAVE-BASED TRANSIENT ANALYSIS USING BLOCK NEWTON-JACOBI
WAVE-BASED TRANSIENT ANALYSIS USING BLOCK NEWTON-JACOBI Muhammad Kabir McGill Uiversity Departmet of Electrical ad Computer Egieerig Motreal, QC H3A 2A7 Email: muhammad.kabir@mail.mcgill.ca Carlos Christofferse
More informationSingle Bit DACs in a Nutshell. Part I DAC Basics
Sigle Bit DACs i a Nutshell Part I DAC Basics By Dave Va Ess, Pricipal Applicatio Egieer, Cypress Semicoductor May embedded applicatios require geeratig aalog outputs uder digital cotrol. It may be a DC
More informationA New Design of Log-Periodic Dipole Array (LPDA) Antenna
Joural of Commuicatio Egieerig, Vol., No., Ja.-Jue 0 67 A New Desig of Log-Periodic Dipole Array (LPDA) Atea Javad Ghalibafa, Seyed Mohammad Hashemi, ad Seyed Hassa Sedighy Departmet of Electrical Egieerig,
More informationAn Adaptive Image Denoising Method based on Thresholding
A Adaptive Image Deoisig Method based o Thresholdig HARI OM AND MANTOSH BISWAS Departmet of Computer Sciece & Egieerig Idia School of Mies, Dhabad Jharkad-86004 INDIA {hariom4idia, matoshb}@gmail.com Abstract
More informationOn Parity based Divide and Conquer Recursive Functions
O Parity based Divide ad Coquer Recursive Fuctios Sug-Hyu Cha Abstract The parity based divide ad coquer recursio trees are itroduced where the sizes of the tree do ot grow mootoically as grows. These
More informationHigh Speed Area Efficient Modulo 2 1
High Speed Area Efficiet Modulo 2 1 1-Soali Sigh (PG Scholar VLSI, RKDF Ist Bhopal M.P) 2- Mr. Maish Trivedi (HOD EC Departmet, RKDF Ist Bhopal M.P) Adder Abstract Modular adder is oe of the key compoets
More informationCh 9 Sequences, Series, and Probability
Ch 9 Sequeces, Series, ad Probability Have you ever bee to a casio ad played blackjack? It is the oly game i the casio that you ca wi based o the Law of large umbers. I the early 1990s a group of math
More informationThe Firing Dispersion of Bullet Test Sample Analysis
Iteratioal Joural of Materials, Mechaics ad Maufacturig, Vol., No., Ma 5 The Firig Dispersio of Bullet Test Sample Aalsis Youliag Xu, Jubi Zhag, Li Ma, ad Yoghai Sha Udisputed, this approach does reduce
More informationTehrani N Journal of Scientific and Engineering Research, 2018, 5(7):1-7
Available olie www.jsaer.com, 2018, 5(7):1-7 Research Article ISSN: 2394-2630 CODEN(USA): JSERBR 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
More information7. Counting Measure. Definitions and Basic Properties
Virtual Laboratories > 0. Foudatios > 1 2 3 4 5 6 7 8 9 7. Coutig Measure Defiitios ad Basic Properties Suppose that S is a fiite set. If A S the the cardiality of A is the umber of elemets i A, ad is
More informationCombined Scheme for Fast PN Code Acquisition
13 th Iteratioal Coferece o AEROSPACE SCIENCES & AVIATION TECHNOLOGY, ASAT- 13, May 6 8, 009, E-Mail: asat@mtc.edu.eg Military Techical College, Kobry Elkobbah, Cairo, Egypt Tel : +(0) 4059 4036138, Fax:
More informationSIDELOBE SUPPRESSION IN OFDM SYSTEMS
SIDELOBE SUPPRESSION IN OFDM SYSTEMS Iva Cosovic Germa Aerospace Ceter (DLR), Ist. of Commuicatios ad Navigatio Oberpfaffehofe, 82234 Wesslig, Germay iva.cosovic@dlr.de Vijayasarathi Jaardhaam Muich Uiversity
More informationOptimal Arrangement of Buoys Observable by Means of Radar
Optimal Arragemet of Buoys Observable by Meas of Radar TOMASZ PRACZYK Istitute of Naval Weapo ad Computer Sciece Polish Naval Academy Śmidowicza 69, 8-03 Gdyia POLAND t.praczy@amw.gdyia.pl Abstract: -
More informationINCREASE OF STRAIN GAGE OUTPUT VOLTAGE SIGNALS ACCURACY USING VIRTUAL INSTRUMENT WITH HARMONIC EXCITATION
XIX IMEKO World Cogress Fudametal ad Applied Metrology September 6, 9, Lisbo, Portugal INCREASE OF STRAIN GAGE OUTPUT VOLTAGE SIGNALS ACCURACY USING VIRTUAL INSTRUMENT WITH HARMONIC EXCITATION Dalibor
More informationDesign of FPGA- Based SPWM Single Phase Full-Bridge Inverter
Desig of FPGA- Based SPWM Sigle Phase Full-Bridge Iverter Afarulrazi Abu Bakar 1, *,Md Zarafi Ahmad 1 ad Farrah Salwai Abdullah 1 1 Faculty of Electrical ad Electroic Egieerig, UTHM *Email:afarul@uthm.edu.my
More informationAnalysis of SDR GNSS Using MATLAB
Iteratioal Joural of Computer Techology ad Electroics Egieerig (IJCTEE) Volume 5, Issue 3, Jue 2015 Aalysis of SDR GNSS Usig MATLAB Abstract This paper explais a software defied radio global avigatio satellite
More information28.3. Kaushik Roy Dept. of ECE, Purdue University W. Lafayette, IN 47907, U. S. A.
8.3 Novel Sizig Algorithm for Yield Improvemet uder Process Variatio i Naometer Techology Seug Hoo Choi Itel Corporatio Hillsboro, OR 974, U. S. A. seug.h.choi@itel.com Bipul C. Paul Dept. of ECE, Purdue
More informationRoberto s Notes on Infinite Series Chapter 1: Series Section 2. Infinite series
Roberto s Notes o Ifiite Series Chapter : Series Sectio Ifiite series What you eed to ow already: What sequeces are. Basic termiology ad otatio for sequeces. What you ca lear here: What a ifiite series
More informationAkinwaJe, A.T., IbharaJu, F.T. and Arogundade, 0.1'. Department of Computer Sciences University of Agriculture, Abeokuta, Nigeria
COMPARATIVE ANALYSIS OF ARTIFICIAL NEURAL NETWORK'S BACK PROPAGATION ALGORITHM TO STATISTICAL LEAST SQURE METHOD IN SECURITY PREDICTION USING NIGERIAN STOCK EXCHANGE MARKET AkiwaJe, A.T., IbharaJu, F.T.
More informationNonlinear System Identification Based on Reduced Complexity Volterra Models Guodong Jin1,a* and Libin Lu1,b
6th Iteratioal Coferece o Electroics, Mechaics, Culture ad Medicie (EMCM 205) Noliear System Idetificatio Based o Reduced Complexity Volterra Models Guodog Ji,a* ad Libi Lu,b Xi a research istitute of
More informationCompound Controller for DC Motor Servo System Based on Inner-Loop Extended State Observer
BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 6, No 5 Special Issue o Applicatio of Advaced Computig ad Simulatio i Iformatio Systems Sofia 06 Prit ISSN: 3-970; Olie ISSN:
More informationSensors & Transducers 2015 by IFSA Publishing, S. L.
Sesors & Trasducers 215 by IFSA Publishig, S. L. http://www.sesorsportal.com Uiversal Sesors ad Trasducers Iterface for Mobile Devices: Metrological Characteristics * Sergey Y. YURISH ad Javier CAÑETE
More informationData Mining of Bayesian Networks to Select Fusion Nodes from Wireless Sensor Networks
www.ijcsi.org http://dx.doi.org/10.20943/01201604.1115 11 Data Miig of Bayesia Networks to Select Fusio Nodes from Wireless Networks Yee Mig Che 1 Chi-Shu Hsueh 2 Chu-Kai Wag 3 1,3 Departmet of Idustrial
More informationOPTIMIZATION OF RNS FIR FILTERS FOR 6-INPUTS LUT BASED FPGAS
OPTIMIZATION OF RNS FIR FILTERS FOR 6-INPUTS LUT BASED FPGAS G.C. Cardarilli, M. Re, A. Salsao Uiversity of Rome Tor Vergata Departmet of Electroic Egieerig Via del Politecico 1 / 00133 / Rome / ITAL {marco.re,
More informationImportance Analysis of Urban Rail Transit Network Station Based on Passenger
Joural of Itelliget Learig Systems ad Applicatios, 201, 5, 22-26 Published Olie November 201 (http://www.scirp.org/joural/jilsa) http://dx.doi.org/10.426/jilsa.201.54027 Importace Aalysis of Urba Rail
More informationDesign and Construction of a Three-phase Digital Energy Meter
Desig ad Costructio of a Three-phase Digital Eergy Meter D.P.Chadima, V.G.R.G. Jayawardae, E.A.E.H. Hemachadra, I.N.Jayasekera, H.V.L.Hasaraga, D.C. Hapuarachchi (chadima@elect.mrt.ac.lk, geethagaj@gmail.com,era.hem@gmail.com,ishaivaka@gmail.com,lahiru_hasaraga@yahoo.com,diya_elect.uom@gmail.com)
More informationLaboratory Exercise 3: Dynamic System Response Laboratory Handout AME 250: Fundamentals of Measurements and Data Analysis
Laboratory Exercise 3: Dyamic System Respose Laboratory Hadout AME 50: Fudametals of Measuremets ad Data Aalysis Prepared by: Matthew Beigto Date exercises to be performed: Deliverables: Part I 1) Usig
More informationThe Institute of Chartered Accountants of Sri Lanka
The Istitute of Chartered Accoutats of Sri Laka Postgraduate Diploma i Busiess ad Fiace Quatitative Techiques for Busiess Hadout 02:Presetatio ad Aalysis of data Presetatio of Data The Stem ad Leaf Display
More informationlecture notes September 2, Sequential Choice
18.310 lecture otes September 2, 2013 Sequetial Choice Lecturer: Michel Goemas 1 A game Cosider the followig game. I have 100 blak cards. I write dow 100 differet umbers o the cards; I ca choose ay umbers
More information8. Combinatorial Structures
Virtual Laboratories > 0. Foudatios > 1 2 3 4 5 6 7 8 9 8. Combiatorial Structures The purpose of this sectio is to study several combiatorial structures that are of basic importace i probability. Permutatios
More informationA Novel Small Signal Power Line Quality Measurement System
IMTC 3 - Istrumetatio ad Measuremet Techology Coferece Vail, CO, USA, - May 3 A ovel Small Sigal Power Lie Quality Measuremet System Paul B. Crilly, Erik Leadro Boaldi, Levy Ely de Lacarda de Oliveira,
More information4. INTERSYMBOL INTERFERENCE
DATA COMMUNICATIONS 59 4. INTERSYMBOL INTERFERENCE 4.1 OBJECT The effects of restricted badwidth i basebad data trasmissio will be studied. Measuremets relative to itersymbol iterferece, usig the eye patter
More informationCross-Layer Performance of a Distributed Real-Time MAC Protocol Supporting Variable Bit Rate Multiclass Services in WPANs
Cross-Layer Performace of a Distributed Real-Time MAC Protocol Supportig Variable Bit Rate Multiclass Services i WPANs David Tug Chog Wog, Jo W. Ma, ad ee Chaig Chua 3 Istitute for Ifocomm Research, Heg
More informationGeneral Model :Algorithms in the Real World. Applications. Block Codes
Geeral Model 5-853:Algorithms i the Real World Error Correctig Codes I Overview Hammig Codes Liear Codes 5-853 Page message (m) coder codeword (c) oisy chael decoder codeword (c ) message or error Errors
More informationReducing Power Dissipation in Complex Digital Filters by using the Quadratic Residue Number System Λ
Reducig Power Dissipatio i Complex Digital Filters by usig the Quadratic Residue Number System Λ Agelo D Amora, Alberto Naarelli, Marco Re ad Gia Carlo Cardarilli Departmet of Electrical Egieerig Uiversity
More informationProblem of calculating time delay between pulse arrivals
America Joural of Egieerig Research (AJER) 5 America Joural of Egieerig Research (AJER) e-issn: 3-847 p-issn : 3-936 Volume-4, Issue-4, pp-3-4 www.ajer.org Research Paper Problem of calculatig time delay
More informationCHAPTER 8 JOINT PAPR REDUCTION AND ICI CANCELLATION IN OFDM SYSTEMS
CHAPTER 8 JOIT PAPR REDUCTIO AD ICI CACELLATIO I OFDM SYSTEMS Itercarrier Iterferece (ICI) is aother major issue i implemetig a OFDM system. As discussed i chapter 3, the OFDM subcarriers are arrowbad
More informationELEC 350 Electronics I Fall 2014
ELEC 350 Electroics I Fall 04 Fial Exam Geeral Iformatio Rough breakdow of topic coverage: 0-5% JT fudametals ad regios of operatio 0-40% MOSFET fudametals biasig ad small-sigal modelig 0-5% iodes (p-juctio
More information(2) The MOSFET. Review of. Learning Outcome. (Metal-Oxide-Semiconductor Field Effect Transistor) 2.0) Field Effect Transistor (FET)
EEEB73 Electroics Aalysis & esig II () Review of The MOSFET (Metal-Oxide-Semicoductor Field Effect Trasistor) Referece: Neame, Chapter 3 ad Chapter 4 Learig Outcome Able to describe ad use the followig:
More informationHigh-Order CCII-Based Mixed-Mode Universal Filter
High-Order CCII-Based Mixed-Mode Uiversal Filter Che-Nog Lee Departmet of Computer ad Commuicatio Egieerig, Taipei Chegshih Uiversity of Sciece ad Techology, Taipei, Taiwa, R. O. C. Abstract This paper
More informationUsing Color Histograms to Recognize People in Real Time Visual Surveillance
Usig Color Histograms to Recogize People i Real Time Visual Surveillace DANIEL WOJTASZEK, ROBERT LAGANIERE S.I.T.E. Uiversity of Ottawa, Ottawa, Otario CANADA daielw@site.uottawa.ca, lagaier@site.uottawa.ca
More informationUnit 5: Estimating with Confidence
Uit 5: Estimatig with Cofidece Sectio 8.2 The Practice of Statistics, 4 th editio For AP* STARNES, YATES, MOORE Uit 5 Estimatig with Cofidece 8.1 8.2 8.3 Cofidece Itervals: The Basics Estimatig a Populatio
More informationMeasurement of Equivalent Input Distortion AN 20
Measuremet of Equivalet Iput Distortio AN 2 Applicatio Note to the R&D SYSTEM Traditioal measuremets of harmoic distortio performed o loudspeakers reveal ot oly the symptoms of the oliearities but also
More informationMethods to Reduce Arc-Flash Hazards
Methods to Reduce Arc-Flash Hazards Exercise: Implemetig Istataeous Settigs for a Maiteace Mode Scheme Below is a oe-lie diagram of a substatio with a mai ad two feeders. Because there is virtually o differece
More informationPulse-echo Ultrasonic NDE of Adhesive Bonds in Automotive Assembly
ECNDT 6 - Poster 7 Pulse-echo Ultrasoic NDE of Adhesive Bods i Automotive Assembly Roma Gr. MAEV, Sergey TITOV, Uiversity of Widsor, Widsor, Caada Abstract. Recetly, adhesive bodig techology has begu to
More informationGeneralization of Selective Harmonic Control/Elimination
Geeralizatio of Selective Harmoic Cotrol/Elimiatio J.R. Wells, P.L. Chapma, P.T. rei Graiger Ceter for Electric Machiery ad Electromechaics Departmet of Electrical ad Computer Egieerig Uiversity of Illiois
More informationAN ESTIMATION OF MULTILEVEL INVERTER FED INDUCTION MOTOR DRIVE
9 IJRIC. All rights reserved. IJRIC www.ijric.org E-ISSN: 76-3336 AN ESTIMATION OF MULTILEVEL INVERTER FED INDUCTION MOTOR DRIVE K.RAMANI AND DR.A. KRISHNAN SMIEEE Seior Lecturer i the Departmet of EEE
More informationA Bipolar Cockcroft-Walton Voltage Multiplier for Gas Lasers
America Joural of Applied cieces 4 (10): 79-799, 007 N 1546-99 007 ciece Publicatios orrespodig Author: A Bipolar ockcroft-walto Voltage Multiplier for Gas Lasers hahid qbal ad Rosli Besar Faculty of Egieerig
More informationProcedia - Social and Behavioral Sciences 128 ( 2014 ) EPC-TKS 2013
Available olie at www.sciecedirect.com ScieceDirect Procedia - Social ad Behavioral Scieces 18 ( 014 ) 399 405 EPC-TKS 013 Iductive derivatio of formulae by a computer Sava Grozdev a *, Veseli Nekov b
More informationVoice Command Recognition System Based on MFCC and VQ Algorithms
Voice Commad Recogitio System Based o MFCC ad VQ Algorithms Mahdi Shaeh, ad Azizollah Taheri Abstract The goal of this project is to desig a system to recogitio voice commads. Most of voice recogitio systems
More informationCancellation of Multiuser Interference due to Carrier Frequency Offsets in Uplink OFDMA
Cacellatio of Multiuser Iterferece due to Carrier Frequecy Offsets i Upli OFDMA S. Maohar, V. Tiiya, D. Sreedhar, ad A. Chocaligam Departmet of ECE, Idia Istitute of Sciece, Bagalore 56001, INDIA Abstract
More informationNovel pseudo random number generation using variant logic framework
Edith Cowa Uiversity Research Olie Iteratioal Cyber Resiliece coferece Cofereces, Symposia ad Campus Evets 011 Novel pseudo radom umber geeratio usig variat logic framework Jeffrey Zheg Yua Uiversity,
More informationFaulty Clock Detection for Crypto Circuits Against Differential Faulty Analysis Attack
Faulty Clock Detectio for Crypto Circuits Agaist Differetial Faulty Aalysis Attack Pei uo ad Yusi Fei Departmet of Electrical ad Computer Egieerig Northeaster Uiversity, Bosto, MA 02115 Abstract. Differetial
More informationSapana P. Dubey. (Department of applied mathematics,piet, Nagpur,India) I. INTRODUCTION
IOSR Joural of Mathematics (IOSR-JM) www.iosrjourals.org COMPETITION IN COMMUNICATION NETWORK: A GAME WITH PENALTY Sapaa P. Dubey (Departmet of applied mathematics,piet, Nagpur,Idia) ABSTRACT : We are
More informationHVIC Technologies for IPM
HVIC Techologies for IPM JONISHI, Akihiro AKAHANE, Masashi YAMAJI, Masaharu ABSTRACT A high voltage itegrated (HVIC), which is a gate driver IC with a high breakdow voltage, is oe of the key devices required
More informationTransfer Functions For The Reference Clock Jitter In A Serial Link: Theory And Applications in PCI Express
DesigCo 25 Leadig Edge Commuicatio Desig Coferece Trasfer Fuctios For The Referece Clock Jitter I A Serial Lik: Theory Ad Applicatios i PCI Express Mike Li, PhD Wavecrest Corporatio 1735 Techology Drive,
More informationCOMPRESSION OF TRANSMULTIPLEXED ACOUSTIC SIGNALS
COMPRESSION OF TRANSMULTIPLEXED ACOUSTIC SIGNALS Mariusz Ziółko, Przemysław Sypka ad Bartosz Ziółko Departmet of Electroics, AGH Uiversity of Sciece ad Techology, al. Mickiewicza 3, 3-59 Kraków, Polad,
More informationReconfigurable architecture of RNS based high speed FIR filter
Idia Joural of Egieerig & Materials Scieces Vol. 21, April 214, pp. 233-24 Recofigurable architecture of RNS based high speed FIR filter J Britto Pari* & S P Joy Vasatha Rai Departmet of Electroics Egieerig,
More informationNovel Modeling Techniques for RTL Power Estimation
Novel Modelig Techiques for RTL Power Estimatio Michael Eierma Walter Stechele Istitute for Itegrated Circuits Istitute for Itegrated Circuits Techical Uiversity of Muich Techical Uiversity of Muich Arcisstr.
More informationEfficient Feedback-Based Scheduling Policies for Chunked Network Codes over Networks with Loss and Delay
Efficiet Feedback-Based Schedulig Policies for Chuked Network Codes over Networks with Loss ad Delay Aoosheh Heidarzadeh ad Amir H. Baihashemi Departmet of Systems ad Computer Egieerig, Carleto Uiversity,
More informationDesign of FPGA Based SPWM Single Phase Inverter
Proceedigs of MUCEET2009 Malaysia Techical Uiversities Coferece o Egieerig ad Techology Jue 20-22, 2009, MS Garde,Kuata, Pahag, Malaysia MUCEET2009 Desig of FPGA Based SPWM Sigle Phase Iverter Afarulrazi
More informationBy: Pinank Shah. Date : 03/22/2006
By: Piak Shah Date : 03/22/2006 What is Strai? What is Strai Gauge? Operatio of Strai Gauge Grid Patters Strai Gauge Istallatio Wheatstoe bridge Istrumetatio Amplifier Embedded system ad Strai Gauge Strai
More informationEfficient Energy Consumption Scheduling: Towards Effective Load Leveling
eergies Article Efficiet Eergy Cosumptio Schedulig: Towards Effective Load Levelig Yua Hog 1, *, Shegbi Wag 2 ad Ziyue Huag 3 1 Departmet of Iformatio Techology Maagemet, Uiversity at Albay, SUNY, 1400
More informationTHE OCCURRENCE OF TRANSIENT FIELDS AND ESD IN TYPICAL SELECTED AREAS
THE OCCURRENCE OF TRANSIENT FIELDS AND ESD IN TYPICAL SELECTED AREAS Stepha FREI Techical Uiversity Berli, Istitute of Electrical Power Egieerig Eisteiufer, 587 Berli, Germay e-mail: frei@ihs.ee.tu-berli.de
More informationDesign of Optimal Hygrothermally Stable Laminates with Extension-Twist Coupling by Ant Colony Optimization
d Iteratioal Coferece o Egieerig Optimizatio September 6-9, 010, Lisbo, Portugal esig of Optimal Hygrothermally Stable Lamiates with Extesio-Twist Couplig by t Coloy Optimizatio ditya P. pte 1, Robert.
More informationSelection of the basic parameters of the lens for the optic-electronic target recognition system
Proceedigs of the 5th WSEAS It. Cof. o COMPUTATIONAL INTELLIGENCE, MAN-MACHINE SYSTEMS AND CYBERNETICS, Veice, Italy, November 0-, 006 317 Selectio of the basic parameters of the les for the optic-electroic
More informationCS 201: Adversary arguments. This handout presents two lower bounds for selection problems using adversary arguments ëknu73,
CS 01 Schlag Jauary 6, 1999 Witer `99 CS 01: Adversary argumets This hadout presets two lower bouds for selectio problems usig adversary argumets ëku73, HS78, FG76ë. I these proofs a imagiary adversary
More informationx 1 + x x n n = x 1 x 2 + x x n n = x 2 x 3 + x x n n = x 3 x 5 + x x n = x n
Sectio 6 7A Samplig Distributio of the Sample Meas To Create a Samplig Distributio of the Sample Meas take every possible sample of size from the distributio of x values ad the fid the mea of each sample
More informationOutline. Motivation. Analog Functional Testing in Mixed-Signal Systems. Motivation and Background. Built-In Self-Test Architecture
Aalog Fuctioal Testig i Mixed-Sigal s Jie Qi Dept. of Electrical & Computer Egieerig Aubur Uiversity Co-Advisors: Charles Stroud ad Foster Dai Outlie Motivatio ad Backgroud Built-I Self-Test Architecture
More informationSHORT-TERM TRAVEL TIME PREDICTION USING A NEURAL NETWORK
SHORT-TERM TRAVEL TIME PREDICTION USING A NEURAL NETWORK Giovai Huiske ad Eric va Berkum Dept. of Civil Egieerig - Uiversity of Twete - 7500 AE Eschede - The Netherlads E-mail: g.huiske@ctw.utwete.l ad
More informationBANDWIDTH AND GAIN ENHANCEMENT OF MULTIBAND FRACTAL ANTENNA BASED ON THE SIERPINSKI CARPET GEOMETRY
ISSN: 2229-6948(ONLINE) DOI: 10.21917/ijct.2013.0095 ICTACT JOURNAL ON COMMUNICATION TECHNOLOGY, MARCH 2013, VOLUME: 04, ISSUE: 01 BANDWIDTH AND GAIN ENHANCEMENT OF MULTIBAND FRACTAL ANTENNA BASED ON THE
More informationAdaptive Resource Allocation in Multiuser OFDM Systems
Adaptive Resource Allocatio i Multiuser OFDM Systems Fial Report Multidimesioal Digital Sigal Processig Malik Meherali Saleh The Uiversity of Texas at Austi malikmsaleh@mail.utexas.edu Sprig 005 Abstract
More informationMaking sure metrics are meaningful
Makig sure metrics are meaigful Some thigs are quatifiable, but ot very useful CPU performace: MHz is ot the same as performace Cameras: Mega-Pixels is ot the same as quality Cosistet ad quatifiable metrics
More informationImplementation of Fuzzy Multiple Objective Decision Making Algorithm in a Heterogeneous Mobile Environment
Implemetatio of Fuzzy Multiple Objective Decisio Makig Algorithm i a Heterogeeous Mobile Eviromet P.M.L. ha, Y.F. Hu, R.E. Sheriff, Departmet of Electroics ad Telecommuicatios Departmet of yberetics, Iteret
More informationH2 Mathematics Pure Mathematics Section A Comprehensive Checklist of Concepts and Skills by Mr Wee Wen Shih. Visit: wenshih.wordpress.
H2 Mathematics Pure Mathematics Sectio A Comprehesive Checklist of Cocepts ad Skills by Mr Wee We Shih Visit: weshih.wordpress.com Updated: Ja 2010 Syllabus topic 1: Fuctios ad graphs 1.1 Checklist o Fuctios
More informationDepartment of Electrical and Computer Engineering, Cornell University. ECE 3150: Microelectronics. Spring Due on April 26, 2018 at 7:00 PM
Departmet of Electrical ad omputer Egieerig, orell Uiersity EE 350: Microelectroics Sprig 08 Homework 0 Due o April 6, 08 at 7:00 PM Suggested Readigs: a) Lecture otes Importat Notes: ) MAKE SURE THAT
More informationFast Sensor Deployment for Fusion-based Target Detection
Fast Sesor Deploymet for Fusio-based Target Detectio Zhaohui Yua*, Rui Ta*, Guoliag Xig*, Cheyag Lu, Yixi Che *Departmet of Computer Sciece, City Uiversity of Hog Kog Departmet of Computer Sciece ad Egieerig,
More information