Observer Gain Adaptation of Output Feedback Sliding Mode Controller with Support Vector Machine Regression

Size: px
Start display at page:

Download "Observer Gain Adaptation of Output Feedback Sliding Mode Controller with Support Vector Machine Regression"

Transcription

1 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso Observer Ga Adaptato of Otpt Feedbac Sldg Mode Cotroller wth Spport Vector Mache Regresso SEZAI OKA SERAR IPLIKCI LUFI ULUSOY Compter Egeerg epartmet Pamale Uverst 7 Kl ezl URKEY Electrcal ad Electrocs Egeerg epartmet Pamale Uverst 7 Kl ezl URKEY {stoat plc}@pa.ed.tr ltflso@hotmal.com Abstract: - he covetoal sldg mode cotroller eeds the eact owledge of sstem state measremets. I ths std olear secod order sstems wth measred sstem states ad boded eteral dstrbaces are cosdered. he sldg mode observer based o olear observato error damcs s cosdered ad the observer ga s adsted b sg a spport vector mache based plat model. From the otpt of the spport vector mache model -step ahead predctos are obtaed. herefore the vale of s frst aalzed to search for a proper vale. It s also show wth the smlatos that the stablt codtos are satsfed for the chose observer gas. Compter smlatos are preseted to show the effect of the proposed ga adstmet mechasm o the performace of otpt feedbac sldg mode cotroller. It s see that the traector tracg performace s mproved wth respect to a covetoal otpt feedbac sldg mode cotrol scheme havg costat sldg mode observer gas. Ke-Words: - Sldg mode cotrol Sldg mode observer Otpt feedbac sldg mode cotrol Spport vector mache regresso Observer ga Boded eteral dstrbaces. Itrodcto Egeers alwas search for better cotrol methods to atta hgher prodctvt tha classcal methods ad to prodce qalt prodcts at compettve prces. Whe ow bt boded eteral dstrbaces are cosdered sldg mode cotrol s a promsg area of std for both theoretcal ad applcato oreted robst cotrol problems []. Sldg mode cotrol s based o varable strctre sstems theor []. It s a olear cotrol method wth a hgh-freqec chatterg pheomeo that alters damcs of a olear sstem. he state-feedbac cotrol law swtches from oe cotos strctre to aother based o the crret posto of sstem traector the state space. he cotrol law mst provde the sstem to move alwas towards a swtchg codto. he moto of the sstem as t sldes alog these bodares s called a sldg mode ad geometrcal locs cosstg of the bodares s called sldg srface []. Sldg mode cotrol strctres have bee appled to varos egeerg problems a wde varet of applcato areas sch as electrcal motors [] moble robots [3] mcro-electromechacal sstems [] chemcal processes ad space sstems. For the mplemetato of the well-ow covetoal sldg mode cotroller SMC strctre eact owledge of sstem state measremets s eeded. However for most applcatos t s ether mpractcal or approprate to se sesors for o-le measremet of all state varables cosderg dfferet reasos as cost relablt harsh evromet or eve dced errors from the sesors [5]. hs ecesst of completel measrg the states of a sstem ca be regarded as a mportat drawbac of covetoal SMCs. Observers ca be sed to replace sesors a cotrol sstem. herefore a cosderable amot of wor has bee doe the feld of state estmato of damc sstems b observers as t s a mportat reqremet for safe ad cost-effectve operato of dstral ts [6]. he observers are frst proposed ad developed for lear sstems [7]. However all practcal sstems heret some degree of oleart ad some cases the lear appromatos based o eact learzato or psedo-learzato ma ot be accrate eogh. ISS: Isse Volme 5 Febrar

2 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso herefore observer theor has bee eteded to clde olear process models [8]. he metoed reqremet for state estmato based o olear observato error damcs ad smple strctre ad robst stablt of SMC prompted the std of sldg mode observers SMOs [9]. I SMOs stead of sg a otpt error feedbac betwee the observer ad sstem learl a olear dscotos term s ected to the observer depedg o the otpt estmato error []. SMOs have a heret robstess the face of eteral dstrbaces ad model certates []. he eqvalet cotrol cocept s proposed [] for lear sstems where the observer states coverge to the sldg srface step b step fte tme. he the eqvalet otpt ecto term that s defed as a coterpart of eqvalet cotrol term of SMC s appled to lear sstems wth ow pts [3]. A SMC that ses the state estmates obtaed from a observer strctre [] or a SMC that ol ses sstem otpts [5] costtte the cocept of otpt feedbac sldg mode cotroller OFSMC. State estmato of olear sstems the presece of eteral dstrbaces or model certates s a actve feld of std [6-8]. he dea [ 3] s eteded [9] to OFSMC desg whch a olear sstem wth ow dstrbaces s cosdered. he parallel processg capabltes of artfcal eral etwor A archtectres provdes a vable meas for costrctg the states of comple damc sstems from pt otpt measremets []. herefore sg soft comptg methodologes order to mprove the performace of SMOs or SMCs s a actve area of research. For stace the speed cotrol of a dcto motor sg a SMC s cosdered [] ad a feedforward A archtectre s sed to estmate the rotor speed. For the SMO case the modelg error of the A observer s compesated b the SMO []. Also a radal bass fcto A ad SMO are sed parallel order to cosder dfferet sstem states or evrometal varables []. I [3] o the other had the A observer ad SMO are coected seral ad the A s sed to obta a olear model of the sstem. For the SMC case a A based observer s sed order to mprove the SMC performace []. For the worst case errors As provde better performace tha lear regresso techqes [5]. However As have a local mma problem whch s a mportat drawbac for most cotrol problems. herefore ths std a spport vector maches SVMs based strctre s chose. SVMs were orgall created to solve classfcato problems. I SMC problems SVMs are sed for dfferet prposes. For stace the desg parameters of a tme-varg sldg srface for a gve tal codto are obtaed b sg SVMs [6]. he the fcto appromato propert of SVMs s sed to desg a ew sldg srface wth addtoal damc states [7]. Also the dscotos cotrol law of SMC s costrcted sg the otpt of SVMs based model order to elmate chatterg [8]. OFSMCs the presece of ow dstrbaces s eamed [9] b provg the stablt der a set of orestrctve assmptos ad t s show that the desged cotroller esres asmptotc traector tracg behavor. o acheve ths am the gas of the state observer mst be properl selected for a acceptable traector tracg performace for the observato error to coverge towards zero. herefore selecto of the observer gas s mportat for the stablt ad performace of the cotroller. A observer that estmates all of the state varables s called a fll-order observer. Whereas a observer that estmates a part of the state varables s referred to be a redced-order observer [9]. I ths std the OFSMC wth a fll-order observer preseted [9] s cosdered ad SVM based plat model ad cotroller tg scheme gve [3] that s developed for tg PI parameters s eteded order to mprove the performace of the SMO ad also to compesate the SMC otpt. he covetoal SMC strateg s orgall desged for cotos-tme operato ad t s more dffclt to choose a sthess for dscrete-tme case [3]. he dscrete-tme SMC s qte dfferet from the covetoal coterpart ad s also called as qas sldg mode. screte-tme SMC desg s sall based o a appromate sldg mode sstem evolto de to the o-qe attractvt codto ad appromate evolto o sldg srface [3]. I or std the parameter adaptato wth SVM s dscrete-tme. However the statefeedbac cotrol scheme based o the SMO SMC ad the plat are all cotos-tme. he strctre of the paper s as follows: I the et secto two ma compoets of the OFSMC scheme the SMC ad SMO are brefl descrbed ad SMC law wth sstem state estmates s preseted. he SVM based modelg predcto ad Jacoba calclatos preseted [3] are gve Secto 3 ad the SMO based ga adaptato scheme s preseted Secto. he smlatos to demostrate the valdt ad advatage of the ga adaptato scheme are gve Secto 5. ISS: Isse Volme 5 Febrar

3 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso Otpt Feedbac Sldg Mode Cotroller he OFSMC cossts of a SMC to geerate the cotrol law ad a SMO to obta the sstem state estmates from measred sstem otpt ad cotrol pt. hese corerstoes of the preseted strctre are emphaszed ths secto.. Cotos-tme Sldg Mode Cotroller he state space represetato of a secod order sgle-pt olear sstem caocal form wth state vector [ ] ca be gve as & & f b t d t where t s the cotrol pt dt s the eteral dstrbaces ad f b are olear fctos that determe the sstem characterstcs [ 9]. he SMC scheme volves selecto of a sldg srface sch that the sstem traector ehbts desrable behavor whe cofed to ths mafold ad fdg feedbac gas so that the sstem traector tersects ad stas o the gve mafold. herefore for sstem assmg the traector tracg problem the error damcs for the secod order sstem gve ca be wrtte as e& & d e& f b t d t & d ad from ths damcs the covetoal lear sldg srface wth costat desg parameters ca be wrtte as s e t c e c e e 3 where e[e e ] s the error state vector ad e s the th error state varable d d s the desred traector of the th state ad c[c ] s the costat sldg srface parameter that determe the sstem behavor the error phase plae. It s ecessar ad sffcet to dfferetate 3 oce for t to appear. hs ths s a frst order stablsato problem based o set. Lapov's drect method cold be sed to obta t that wold eep set at zero. Cosder a Lapov fcto caddate as V s s t wth V Vs> for st> []. A effcet codto for sstem stablt ca be gve as d V& s s t η s t 5 dt where η s a strctl postve real costat that determes the covergece veloct of the traector to the sldg srface. Obtag the eqalt 5 meas that the dstace to the srface decreases alog all traectores ad ths meas that the sstem s stable. herefore 5 s called as the reachablt codto for the sldg srface. B sbstttg 3 to 5 ad omttg the argmets of the depedet varables oe obtas s. f b. b. d& c e& η s 6 d herefore a cotrol pt satsfg the reachg codto ca be chose as f & c e& d b sg s 7 where g s a strctl postve real costat wth a lower bod depedg o the boded eteral dstrbaces. he fcto sg. deotes the sgm fcto defed as follows g f s < sg s f s 8 f s > he cotrol pt 7 cossts of two parts. he frst part eq s the cotos term that s ow as eqvalet cotrol based o estmated sstem parameters ad t compesates the estmated desrable damcs of the sstem. he secod part wth the sgm fcto s the dscotos cotrol law ds that reqres fte swtchg o the part of the cotrol sgal ad actator at the tersecto of error state traector ad sldg srface. I ths wa the traector s forced to move alwas towards the sldg srface [].. Sldg Mode Observer he state estmato problem for a sstem sbect to ow eteral dstrbaces der otpt feedbac sldg mode cotrol wth a eqvalet otpt ecto sldg mode observer s cosdered [9]. I ths std the sldg mode observer strctre preseted [9] s sed. For the sstem gve ol the sstem otpt s measred. herefore the error damcs cold ot be obtaed. he sstem states ad also the error damcs ca be obtaed from b sg a observer of the form gve as & λ sg & f b E λ sg ~ eq ds 9 ISS: Isse Volme 5 Febrar

4 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso where ~ λ sg ad the eqvalet otpt ecto term eq λ sg eq s obtaed b sg a low pass flter [5 9]. he term E f ad E otherwse [9]. Wth proper λ λ observer gas the observer state frstl coverges to ad the coverges to. For the gve sstem fte tme covergece of sstem state estmates to actal plat states s proved the lteratre [9]. herefore stead of 3 obtaed from oe ca se s e t c e c e e where e [ e e ] s estmated error state vector ad e d s th estmated error state varable. If the sstem states are ot measrable the covetoal form of eq sg state estmates ca be rewrtte as t b f & c e & eq d he the overall cotrol law based o estmated sstem states ca be desged as t t sg s eq Choosg the Lapov fcto caddate as V/ ŝ sg estmated state varables ad tag the dervatve of the Lapov fcto alog the traectores of the estmated sstem states the dscotos cotrol ga g mst be chose as [9] g g b c λ λ η 3 order to satsf the reachg codto. 3 Spport Vector Mache based Modelg Predcto ad Jacoba Calclatos Cosder a olear sstem damcs of whch ca be represeted b ARX model f L L where s the cotrol sgal appled to the plat at tme de s the correspodg otpt of the plat ad ad deote the mber of past cotrol sgals ad mber of the past otpts volved the model respectvel. It s assmed that o-lear fcto f s ow ad that a trag data set s obtaed the form gve as { L set { } L } 5 where X s the th pt data pot pt space ad Y s the correspodg otpt vale. It s desred to obta a model that represets the relatoshp betwee pt ad otpt data pots. he trag data set set s to be sed to obta a appromate model of the plat damcs. he prmal form of a SVM regresso model s gve b 6 whch s lear a hgherdmesoal featre space F w Φ bas 6 where w s a vector the featre space F Φ s a mappg from the pt space to the featre space bas s the bas term ad <> stads for er prodct operato F [3]. he SVR algorthms regard the regresso problem as a optmzato problem dal space whch the model s gve b bas α K 7 where α s are the coeffcets of each trag data ad K s the erel fcto gve b K Φ Φ K [3] he erel fcto K hadles er prodct featre space ad hece the eplct form of Φ does ot eed to be ow. I the model 7 a trag pot correspodg to a ozero α vale s referred to as the spport vector. I [3] ε-svr algorthm emplog Vap s ε- sestve loss fcto L ε gve as ε L ε 8 > ε s sed whch formlates the prmal form of the regresso problem as follows: m P w C 9 ε wbasξ ξ sbect to the costrats ξ ξ ι wφ τ ε ξ w Φ τ ε ξ ξ ξ... where ε s the pper vale of tolerable error ξ are slac varables. s the Ecldea orm ad C s a reglarzato parameter that provdes a compromse betwee model complet ad degree ξ ISS: Isse Volme 5 Febrar

5 of tolerace to the errors larger tha ε [3]. al form of the optmzato problem becomes a QP problem as K m ε ε 3 sbect to the costrats C... solto of the QP problem 3 ad gves the optmm vales of ad s. he vale of bas the model s determed as follows: the codto ε s satsfed for each spport vector for whch the codto C holds. If α s defed to be the ew coeffcet of for as α the a SVM model as gve b 7 s obtaed. Moreover whe ol spport vectors are cosdered the model becomes K bas #SV SV α 5 where #SV s the mber of spport vectors. If we follow the procedre gve [3] the we costrct the crret state vector as ] [ L L v 6 the the correspodg otpt of the SVM model becomes K bas #SV SV v α 7 I SVM-based observer ga adaptato a radal bass adopted erel fcto s sed that s gve as - - ep K K 8 where s the wdth parameter [3]. If s defed as Ecldea dstace betwee th spport vector ad crret state vector v as v v 9 the the erel fcto ca be rewrtte as ep K v 3 ad the SVM regresso model becomes bas #SV ep α 3 ow 3 ca be sed to predct -step ahead ftre traector of the plat as b K bas... ep #SV α 3 where < m 33 the frst order partal dervatves ca be wrtte as SV # ep α 3 where ep ep ep d 35 ad ] [ ] [ m δ δ 36 where. δ stads for t step fcto [3]. ow the frst-order terms ca be sed to calclate the Jacoba matr. WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso ISS: Isse Volme 5 Febrar

6 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso. Spport Vector Mache Based Observer Ga Adaptato he proposed SVM based sldg mode observer ga adaptato scheme s gve Fg.. It s adopted from the std proposed [3] whch s frst sed for tg PI cotroller parameters. he dea s mal based o obtag the -step ahead predctos of the plat otpt b sg a SVM model ad a Jacoba bloc for tg SMO gas. he SVM model s obtaed b applg radoml chose boded cotrol sgals to the plat. After the trag process -step ahead predctos [... ] are obtaed from the otpt of the SVM model wth t d tme dratos as show Fg.. he order to mmze the SVM predcto error ad to pealze the wated rapd chages the cotrol pt a obectve fcto s chose as [3] φ ε ρ 37 d where ε s the predcto error of SVM at th d step s the ow desred otpt at th step ad ρ determes the amot of pealt o the cotrol devatos. I ths std the proposed dea [3] s appled to the OFSMC case b choosg the observer ga λ as the pdated parameter. I order to have a mercal solto to the problem of mmzg 37 Leveberg- Marqardt learg rle whch terpolate betwee Gass-ewto ad steepest descet algorthms ca be wrtte as J J I J ε λ λ κ 38 ew old where κ s a bledg factor whch determes a mg rato betwee gradet-descet ad Gass- ewto algorthms ad ε s the predcto error vector whch s defed as ε [ ε ε L ε ρ ] 39 he plat s desred otpt traector does ot deped o observer gas. herefore Jacoba matr J 38 ca be obtaed from 37 as ρ J L λ λ λ λ Usg the preseted scheme the observer ga λ s pdated dscretel at ever samplg perod t d ad sed to pdate the SMO that s cotos tme. AC SMO Eq.9 Fgre. Schematc dagram of the SVM based observer ga adaptato ad cotrol law compesato scheme. he cotrol law compesato mechasm ca also be obtaed b splttg the Jacoba matr to two dfferet parts b sg secod order alor appromato of 37 [3]. hs applg the gve dea to the observer ga adaptato scheme the Jacoba matr ca be wrtte as follows J ρ λ he partal dervatve of wth respect λ cold be drectl obtaed b solvg the eqatos from the SMO ad SMC blocs whch have both olear strctres. hs olear strctre rases dffcltes obtag the mathematcal solto. herefore ths std the mercal soltos are obtaed b sg the appromato gve as λ λ From the stablt aalss gve [9] order to provde the fte tme covergece of the estmated states to the actal states the observer gas mst satsf the codtos gve as λ > J Eq. SMC Eq. µ d Otpt Feedbac SMC Scheme t d ŷ SVM Model Eq.3 Plat Eq. λ > f b d f b µ û AC 3 where µ µ are small postve real costats [9]. herefore these bods mst be provded whe tg the λ observer ga. Itall λ s set to acceptable vales that provde 3. Proper choce of the gas λ ad λ wll garatee that the redced order damcs are stable o the sldg srface ad t d d ISS: Isse Volme 5 Febrar

7 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso ths wll esre asmptotc stablt of the referece traector. o have a better observer performace ad ths to provde a better otpt tracg performace ths preset vale shold be ted properl. 5 Smlato Stdes o show the performace of the ew tg scheme compter smlatos are performed o a olear mass-sprg-damper sstem o a horzotal srface der the effect of a horzotal force. he damc eqatos of the sstem s descrbed as m && v & t t t d t v & t v & v & & 3 t where m s the mass t s the dsplacemet & t s the veloct v & t ad t are olear terms wth respect to the damper ad sprg respectvel. B tag & ad b rewrtg the sstem eqatos the form of oe ca obta f v & t t t d t m b / m 5 he sstem parameters are chose as m v o v.35 ad o.55. he tal state vales are chose as.5. he traector tracg problem s cosdered ad the desred state traectores are chose as t.5cos πt/5 d t.π s πt/5 d 6 rg the smlatos order to show the robstess agast boded eteral dstrbaces dt s modeled wth a ssodal sgal tae as d t.5.5cos3πt 7 he SMO for all OFSMCs s tae as 9 ad to obta ~ frst order low pass flter wth badwdth w rad/s s sed. For all of the cotrollers the sldg srface parameter s tae as c 7. Smlatos have bee carred ot Matlab evromet ad ordar dfferetal eqato solver mplemetg Rge-Ktta mercal tegrato method has bee selected for smlatg the dscotos atre of sldg mode cotroller ad observer. For the smlato evromet a fed samplg tme of e-s has bee appled for smlatg the cotos tme observer cotroller ad plat. O the other had the SVM bloc wors a dscrete atre b tag observatos ad calclatg the pdate vale at ever t d e- s tme dratos. All smlatos are performed the tme terval betwee [ 5] s. he sstem performace s fleced b the selecto of the observer tal codtos. herefore assmg that the tal vales of sstem states ad are at the org average ad. he SVM predcts -step ahead sstem behavor ad s a desg parameter. o aalze the effect of o the performace dces ad cotrol pt magtde the sstem s smlated for dfferet vales of betwee [:] ad the reslts are gve Fg.-3. As ca be see from Fg. the performace has ts best vales for ad the performace s the smlar for. However from Fg.3 t s see that for ths performace mprovemet has a trade-off as a creased cotrol pt magtde. herefore ad are chose for comparso. he traector tracg ad state estmato performaces are gve Fg. ad the cotrol pts are gve Fg.5 for ad respectvel. ta ts treachs treachŝ b a c d Fgre. Performace dces for dfferet vales. ISS: Isse Volme 5 Febrar

8 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso ma Fgre 3. Cotrol pt magtde ma. Fgre. Actal estmated ad referece sstem otpt for a b d a.5 d b tme [s] - - Fgre 5. Cotrol pts for a a b tme [s] b a.8.6 λ m λ b tme [s] Fgre 6. Observer gas ad ther stablt bods 3 for : a λ m λ b λ m λ. λ m λ.5.5 Fgre 7. Observer gas ad ther stablt bods 3 for : a λ m λ b λ m λ. he performace for Fg. s better tha the case for. However the cotrol pt magtde Fg.5 reaches ear 6 for whch s far more tha the covetoal case. For the stablt of the sstem the codtos gve 3 mst be provded. he λ λ vales ad ther mmm vales that are obtaed from 3 are plotted Fg.6 ad 7 for ad respectvel. It s see that the observer gas do ot cocde wth the gve stablt codtos. Cosderg above aalss o sstem stablt ad performace for detaled comparsos cotrol pt magtde s also cosdered ad for the -step ahead predcto otpt of the SVM model s chose. he smlatos are mplemeted for the proposed OFSMC wth SVM based ga adaptato scheme OFSMC-SVM ad for the covetoal OFSMC preseted [9] OFSMC-C. Bearg md the stablt codtos [9] three dfferet cases for OFSMC-C s cosdered: OFSMC-C : λ. λ. 8 ad g 3.6 OFSMC-C : λ. λ. 8 ad g OFSMC-C 3 : λ. 93 λ. 8 ad g he three cases for OFSMC-C are desged order to show the effect of adsted vales obtaed b the SVM scheme. For OFSMC-C OFSMC-C 3 ad OFSMC-SVM g s calclated from 3 wth µ.. For OFSMC-C o the other had g s chose as the mamm vale obtaed wth OFSMC-SVM. OFSMC-C 3 has costat λ vale that s obtaed at last wth OFSMC-SVM ad g s calclated from 3 for costat λ.93. he tal vale of λ. s chose for OFSMC-SVM. he bodares for λ s tae as λ m. λ m λ a.8.6 λ. m λ b tme [s] ISS: Isse Volme 5 Febrar

9 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso ad λ ma 5. he bledg factor 38 s tae as κ. ad the pealt o the cotrol devatos s ρ.. he tme resposes of estmated sstem otpt ad actal sstem otpt are gve Fg.8. For all the cotrollers there s some traset observato error at the begg of observato as observer tal codtos are cosstet wth those of the plat. Bt observer states ad ths sstem otpt estmate approach to ts actal vale after a fte tme. he sldg srface s ad estmated sldg srface ŝ are plotted Fg.9. he sldg moto Fg.9 provdes a estmate of the sstem states. For OFSMC-C creasg g mproves t reach ŝ. However ths does ot mprove the observer behavor as ca be easl see from the vale of t reach s. he cotrol pts are also plotted Fg. ad chatterg s a reslt of sgm fcto ad ca be avoded b sg a satrato fcto. he tme-varg behavor of the pdated λ t ad calclated g t for the proposed OFSMC-SVM cotroller s plotted Fg.. he g s calclated from 3 b sg the tme-varg λ vale whch s calclated wth the proposed method at each t d tme tervals. At tme t.57 s the parameters reach ther optmm vales ad sta costat as λ.93 ad g after that tme stat..5 d -.5 a.5 d -.5 b.5 d -.5 c.5 d -.5 d tme [s] Fgre 8. Actal estmated ad referece sstem otpt : aofsmc-c bofsmc-c cofsmc-c 3 dofsmc-svm. a Fgre 9. Actal ad estmated sldg srface varables: aofsmc-c bofsmc-c cofsmc- C 3 dofsmc-svm. 6 6 b 6 c 6 d - a - b - c - d Fgre. Cotrol pts: a OFSMC-C b OFSMC-C c OFSMC-C 3 d OFSMC-SVM. λ a 8 g 6 tme [s] tme [s] b tme [s] Fgre. a λ ad b g for OFSMC-SVM s ŝ s ŝ s ŝ s ŝ ISS: Isse Volme 5 Febrar

10 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso able. Performace dces of the cotrollers. OFSMC- OFSMC- OFSMC- OFSMC- C C C 3 SVM t a t a t s t s t reach s t reach ŝ ma{ } Ol oe of the observer gas s cosdered ths std. However the observer ga adstmet mechasm preseted ths std ca be appled to both observer gas. Also the case of hgher order sstems the form of the gve strctre proposed method ca be eteded b cosderg the stablt codtos. Acowledgemet hs pblcato s a reslt of a proect carred ot over the perod 8-. he athors wsh to tha he Scetfc ad echologcal Research Cocl of re UBIAK for facal spport Proect o: 7E86. Fall the performace dces of the related cotrollers are gve able. he error bod for the settlg tme s tae as 5% of the stead state vale. I able t a s the tme that estmated state approach ts actal vale t s s the settlg tme for state ad t reach s the reachg tme of estmated ad actal sldg srface varables. It s see that ol creasg g does ot have a postve effect o t a. he OFSMC-C 3 represets the obtaed vales geerated b sg OFSMC-SVM adaptato scheme. hs t has the best observato behavor ad settlg tme performace whch shows the costrctve tg strateg of the preseted SVM model. 6 Coclso I ths std otpt feedbac sldg mode cotrol of a olear secod order sstem sbect to boded eteral dstrbaces s cosdered. he ovelt of ths std s that the spport vector mache regresso algorthm s frstl sed wth the otpt feedbac sldg mode cotroller strctre. I ths std the spport vector mache regresso algorthm s sed to adst the sldg mode observer gas. B sg compter smlatos t s see that the mber of ftre data pots predcted b the spport vector mache based model flece both the performace of the sstem ad the magtde of the cotrol pt. herefore a proper vale for the mber of ftre data pots s selected. From the smlato reslts t was coclded that the proposed method mproves the sstem traector tracg performace ad the observer gas. Also t s show wth the smlatos that the stablt codtos for the observer gas are satsfed. Refereces: [] A.S.I. Zober A trodcto to sldg mode varable strctre cotrol I Varable Strctre ad Lapov Cotrol Zober A.S.I. ed. Lectre otes Cotrol ad Iformato Sceces Vol.93/99 Sprger Verlag Lodo 99 pp [] O. Baramboes F.J. Maseda A.J Garrdo ad P. Gomez A Sldg Mode Cotrol Scheme for Idcto Motors Usg eral etwors for Rotor Speed Estmato Proceedgs of the 5 th WSEAS Iteratoal Coferece o Istrmetato Measremet Cotrol Crcts ad Sstems Cac Meco Ma - 5 pp [3] A. Flpesc A.L. Stac S. Flpesc ad G. Stamatesc O-le Parameter Estmato Sldg-mode Cotrol of Poeer 3-X Wheeled Moble Robot Proceedgs of the 7 th WSEAS Iteratoal Coferece o Sstems heor ad Scetfc Comptato Athes Greece Agst -6 7 pp [] A. Kz S. Bogosa ad M. Goasa Cotrol Strateges for Icreased Relablt MEM Comb rves Proceedgs of the d WSEAS Iteratoal Coferece o Appled ad heoretcal Mechacs Vece Ital ovember - 6 pp.3-6. [5] I. Hasara U. Ozger ad V. Ut O sldg mode observers va eqvalet cotrol approach Iteratoal Joral of Cotrol Vol.7 o pp [6] G. Ells Observers cotrol sstems: a practcal gde Academc Press Sa ego CA. [7].G. Leberger A trodcto to observers IEEE rasactos o Atomatc Cotrol Vol.AC-6 97 pp ISS: Isse Volme 5 Febrar

11 WSEAS RASACIOS o SYSEMS ad COROL Seza oat Serdar Iplc Ltf Ulso [8] F.E. ha Observg the state of olear damcal sstems Iteratoal Joral of Cotrol Vol pp [9] A. McCa M.S. Islam ad I. Hssa Applcato of a sldg-mode observer for posto ad speed estmato swtched relctace motor drves IEEE rasactos o Idstr Applcatos Vol.37 o. pp [] C. Edwards S.K. Sprgeo C.P. a ad. Patel Sldg mode observers Lectre otes Cotrol & Iformato Sstems Vol pp.-. [] S.K. Sprgeo Sldg mode observers: a srve Iteratoal Joral of Sstems Scece Vol.39 o.8 8 pp [] S. raov ad V. Ut Sldg mode observers: a ttoral Proceedgs of the 3th IEEE Iteratoal Coferece o ecso ad Cotrol 995 pp [3]. Floqet ad J.-P. Barbot A caocal form for the desg of ow pt sldg mode observers I Advaces Varable Strctre ad Sldg Mode Cotrol C. Edwards C.E. Fossas L. Frdma eds. Lectre otes Cotrol ad Iformato Sceces Vol.33 Sprger Berl 6 pp [] C. Usal ad P. Kachroo Sldg mode measremet feedbac cotrol for atloc brag sstem IEEE rasactos o Cotrol Sstems echolog Vol.7 o. 999 pp.7-8. [5] M.C. Pa screte-tme otpt feedbac sldg mode cotrol for certa sstems Joral of Mare Scece ad echolog Vol.6 o. 8 pp [6] E.H.E. Baom Speed sesor-less sldg mode cotrol of dcto motor drve WSEAS rasactos o Crcts ad Sstems Vol.3 o. 8 pp [7] W. Sagtgtog ad S. Stor Adaptve Sldg-Mode Speed-orqe Observer. WSEAS rasactos o Sstems Vol.5 o.3 6 pp [8] W. Sagtgtog ad S. Stor Stablt Aalss of a Sldg-Mode Speed Observer drg raset State Proceedgs of the 5 th WSEAS Iteratoal Coferece o Istrmetato Measremet Crcts ad Sstems Hagzho Cha Aprl pp.35-. [9] J.M. al ad.w.l. Wag Otpt feedbac sldg mode cotrol the presece of ow dstrbaces Sstems ad Cotrol Letters Vol.58 9 pp [] A. Baz A eral observer for damc sstems Joral of Sod ad Vbrato Vol.5 o. 99 pp.7-3. [] I. Charez A. Poza ad. Poza ew sldg-mode learg law for damc eral etwor observer IEEE rasactos o Crcts ad Sstems II Epress Brefs Vol.53 o. 6 pp [] W. Y Stablt aalss of vsal servog wth sldg mode estmato ad eral comptato Iteratoal Joral of Cotrol Atomato ad Sstems Vol. o.5 6 pp [3] J. Resedz W. Y ad L. Frdma wo-stage eral observer for mechacal sstems IEEE rasactos o Crcts ad Sstems II Epress Brefs Vol.55 o. 8 pp [] M. Lee A sldg mode cotroller wth eral etwor ad fzz logc Proceedgs of the IEEE Iteratoal Coferece o eral etwors Vol pp.-7. [5] V. Krova ad M. Saget Comparso of worst case errors lear ad eral etwor appromato IEEE rasactos o Iformato heor Vol.8 o. pp [6] S. oat me-varg sldg srface desg wth spport vector mache based tal codto adaptato Joral of Vbrato ad Cotrol Vol. o.8 6 pp [7] F.G. Wag S.K. Par M.C. Km S.J. Cho ad.s. Yoo A ovel sldg srface desg b sg spport vector maches Proceedgs of the Iteratoal Coferece o Comple Sstems ad Applcatos Ja Cha 7 pp.5-9. [8] J.. L Y.B. Zhag ad H.P. Pa Chatterg free LS-SVM Sldg Mode Cotrol Proceedgs of the 5 th Iteratoal Smposm o eral etwors Beg Cha 8 pp [9] E.M. Jafarov A ew Redced-order sldg mode observer desg method: A trple trasformatos approach Proceedgs of the 9 th WSEAS Iteratoal Coferece o Sstems Athes Greece Jl -3 5 pp.-8. [3] S. Iplc A comparatve std o a ovel model-based PI tg ad cotrol mechasm for olear sstems Iteratoal Joral of Robst ad olear Cotrol do../rc.5 9. ISS: Isse Volme 5 Febrar

DISTURBANCE TERMS. is a scalar and x i

DISTURBANCE TERMS. is a scalar and x i DISTURBANCE TERMS I a feld of research desg, we ofte have the qesto abot whether there s a relatoshp betwee a observed varable (sa, ) ad the other observed varables (sa, x ). To aswer the qesto, we ma

More information

2.160 System Identification, Estimation, and Learning Lecture Notes No. 17 April 24, 2006

2.160 System Identification, Estimation, and Learning Lecture Notes No. 17 April 24, 2006 .6 System Idetfcato, Estmato, ad Learg Lectre Notes No. 7 Aprl 4, 6. Iformatve Expermets. Persstece of Exctato Iformatve data sets are closely related to Persstece of Exctato, a mportat cocept sed adaptve

More information

B-spline curves. 1. Properties of the B-spline curve. control of the curve shape as opposed to global control by using a special set of blending

B-spline curves. 1. Properties of the B-spline curve. control of the curve shape as opposed to global control by using a special set of blending B-sple crve Copyrght@, YZU Optmal Desg Laboratory. All rghts reserved. Last pdated: Yeh-Lag Hs (--9). ote: Ths s the corse materal for ME Geometrc modelg ad compter graphcs, Ya Ze Uversty. art of ths materal

More information

CS 1675 Introduction to Machine Learning Lecture 12 Support vector machines

CS 1675 Introduction to Machine Learning Lecture 12 Support vector machines CS 675 Itroducto to Mache Learg Lecture Support vector maches Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Mdterm eam October 9, 7 I-class eam Closed book Stud materal: Lecture otes Correspodg chapters

More information

Lecture 2: The Simple Regression Model

Lecture 2: The Simple Regression Model Lectre Notes o Advaced coometrcs Lectre : The Smple Regresso Model Takash Yamao Fall Semester 5 I ths lectre we revew the smple bvarate lear regresso model. We focs o statstcal assmptos to obta based estmators.

More information

Binary classification: Support Vector Machines

Binary classification: Support Vector Machines CS 57 Itroducto to AI Lecture 6 Bar classfcato: Support Vector Maches Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square CS 57 Itro to AI Supervsed learg Data: D { D, D,.., D} a set of eamples D, (,,,,,

More information

Processing of Information with Uncertain Boundaries Fuzzy Sets and Vague Sets

Processing of Information with Uncertain Boundaries Fuzzy Sets and Vague Sets Processg of Iformato wth Ucerta odares Fzzy Sets ad Vage Sets JIUCHENG XU JUNYI SHEN School of Electroc ad Iformato Egeerg X'a Jaotog Uversty X'a 70049 PRCHIN bstract: - I the paper we aalyze the relatoshps

More information

Support vector machines II

Support vector machines II CS 75 Mache Learg Lecture Support vector maches II Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Learl separable classes Learl separable classes: here s a hperplae that separates trag staces th o error

More information

( ) = ( ) ( ) Chapter 13 Asymptotic Theory and Stochastic Regressors. Stochastic regressors model

( ) = ( ) ( ) Chapter 13 Asymptotic Theory and Stochastic Regressors. Stochastic regressors model Chapter 3 Asmptotc Theor ad Stochastc Regressors The ature of eplaator varable s assumed to be o-stochastc or fed repeated samples a regresso aalss Such a assumpto s approprate for those epermets whch

More information

( ) ( ) ( ( )) ( ) ( ) ( ) ( ) ( ) = ( ) ( ) + ( ) ( ) = ( ( )) ( ) + ( ( )) ( ) Review. Second Derivatives for f : y R. Let A be an m n matrix.

( ) ( ) ( ( )) ( ) ( ) ( ) ( ) ( ) = ( ) ( ) + ( ) ( ) = ( ( )) ( ) + ( ( )) ( ) Review. Second Derivatives for f : y R. Let A be an m n matrix. Revew + v, + y = v, + v, + y, + y, Cato! v, + y, + v, + y geeral Let A be a atr Let f,g : Ω R ( ) ( ) R y R Ω R h( ) f ( ) g ( ) ( ) ( ) ( ( )) ( ) dh = f dg + g df A, y y A Ay = = r= c= =, : Ω R he Proof

More information

Regression and the LMS Algorithm

Regression and the LMS Algorithm CSE 556: Itroducto to Neural Netorks Regresso ad the LMS Algorthm CSE 556: Regresso 1 Problem statemet CSE 556: Regresso Lear regresso th oe varable Gve a set of N pars of data {, d }, appromate d b a

More information

Functions of Random Variables

Functions of Random Variables Fuctos of Radom Varables Chapter Fve Fuctos of Radom Varables 5. Itroducto A geeral egeerg aalyss model s show Fg. 5.. The model output (respose) cotas the performaces of a system or product, such as weght,

More information

Chapter 11 Systematic Sampling

Chapter 11 Systematic Sampling Chapter stematc amplg The sstematc samplg techue s operatoall more coveet tha the smple radom samplg. It also esures at the same tme that each ut has eual probablt of cluso the sample. I ths method of

More information

LINEAR EQUALIZERS & NONLINEAR EQUALIZERS. Prepared by Deepa.T, Asst.Prof. /TCE

LINEAR EQUALIZERS & NONLINEAR EQUALIZERS. Prepared by Deepa.T, Asst.Prof. /TCE LINEAR EQUALIZERS & NONLINEAR EQUALIZERS Prepared by Deepa.T, Asst.Prof. /TCE Eqalzers The goal of eqalzers s to elmate tersymbol terferece (ISI) ad the addtve ose as mch as possble. Itersymbol terferece(isi)

More information

An Expansion of the Derivation of the Spline Smoothing Theory Alan Kaylor Cline

An Expansion of the Derivation of the Spline Smoothing Theory Alan Kaylor Cline A Epaso of the Derato of the Sple Smoothg heory Ala Kaylor Cle he classc paper "Smoothg by Sple Fctos", Nmersche Mathematk 0, 77-83 967) by Chrsta Resch showed that atral cbc sples were the soltos to a

More information

Comparison of Dual to Ratio-Cum-Product Estimators of Population Mean

Comparison of Dual to Ratio-Cum-Product Estimators of Population Mean Research Joural of Mathematcal ad Statstcal Sceces ISS 30 6047 Vol. 1(), 5-1, ovember (013) Res. J. Mathematcal ad Statstcal Sc. Comparso of Dual to Rato-Cum-Product Estmators of Populato Mea Abstract

More information

Support vector machines

Support vector machines CS 75 Mache Learg Lecture Support vector maches Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square CS 75 Mache Learg Outle Outle: Algorthms for lear decso boudary Support vector maches Mamum marg hyperplae.

More information

CS 2750 Machine Learning. Lecture 8. Linear regression. CS 2750 Machine Learning. Linear regression. is a linear combination of input components x

CS 2750 Machine Learning. Lecture 8. Linear regression. CS 2750 Machine Learning. Linear regression. is a linear combination of input components x CS 75 Mache Learg Lecture 8 Lear regresso Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square CS 75 Mache Learg Lear regresso Fucto f : X Y s a lear combato of put compoets f + + + K d d K k - parameters

More information

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model Lecture 7. Cofdece Itervals ad Hypothess Tests the Smple CLR Model I lecture 6 we troduced the Classcal Lear Regresso (CLR) model that s the radom expermet of whch the data Y,,, K, are the outcomes. The

More information

Kernel-based Methods and Support Vector Machines

Kernel-based Methods and Support Vector Machines Kerel-based Methods ad Support Vector Maches Larr Holder CptS 570 Mache Learg School of Electrcal Egeerg ad Computer Scece Washgto State Uverst Refereces Muller et al. A Itroducto to Kerel-Based Learg

More information

A Robust Total Least Mean Square Algorithm For Nonlinear Adaptive Filter

A Robust Total Least Mean Square Algorithm For Nonlinear Adaptive Filter A Robust otal east Mea Square Algorthm For Nolear Adaptve Flter Ruxua We School of Electroc ad Iformato Egeerg X'a Jaotog Uversty X'a 70049, P.R. Cha rxwe@chare.com Chogzhao Ha, azhe u School of Electroc

More information

Motion Estimation Based on Unit Quaternion Decomposition of the Rotation Matrix

Motion Estimation Based on Unit Quaternion Decomposition of the Rotation Matrix Moto Estmato Based o Ut Qatero Decomposto of the Rotato Matrx Hag Y Ya Baozog (Isttte of Iformato Scece orther Jaotog Uversty Bejg 00044 PR Cha Abstract Based o the t qatero decomposto of rotato matrx

More information

1 Lyapunov Stability Theory

1 Lyapunov Stability Theory Lyapuov Stablty heory I ths secto we cosder proofs of stablty of equlbra of autoomous systems. hs s stadard theory for olear systems, ad oe of the most mportat tools the aalyss of olear systems. It may

More information

Objectives of Multiple Regression

Objectives of Multiple Regression Obectves of Multple Regresso Establsh the lear equato that best predcts values of a depedet varable Y usg more tha oe eplaator varable from a large set of potetal predctors {,,... k }. Fd that subset of

More information

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971))

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971)) art 4b Asymptotc Results for MRR usg RESS Recall that the RESS statstc s a specal type of cross valdato procedure (see Alle (97)) partcular to the regresso problem ad volves fdg Y $,, the estmate at the

More information

The Double Rotation CORDIC Algorithm: New Results for VLSI Implementation of Fast Sine/Cosine Generation

The Double Rotation CORDIC Algorithm: New Results for VLSI Implementation of Fast Sine/Cosine Generation he Doble Rotato CORDIC Algorthm: New Reslts for VLSI Implemetato of Fast Se/Cose eerato ze-y Sg * Chch-S Che ** Mg-Cho Shh * * Departmet of Electrcal Egeerg ** Isttte of Egeerg Scece Chg Ha Uerst, Hsch,

More information

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines It J Cotemp Math Sceces, Vol 5, 2010, o 19, 921-929 Solvg Costraed Flow-Shop Schedulg Problems wth Three Maches P Pada ad P Rajedra Departmet of Mathematcs, School of Advaced Sceces, VIT Uversty, Vellore-632

More information

Supervised learning: Linear regression Logistic regression

Supervised learning: Linear regression Logistic regression CS 57 Itroducto to AI Lecture 4 Supervsed learg: Lear regresso Logstc regresso Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square CS 57 Itro to AI Data: D { D D.. D D Supervsed learg d a set of eamples s

More information

Chapter Two. An Introduction to Regression ( )

Chapter Two. An Introduction to Regression ( ) ubject: A Itroducto to Regresso Frst tage Chapter Two A Itroducto to Regresso (018-019) 1 pg. ubject: A Itroducto to Regresso Frst tage A Itroducto to Regresso Regresso aalss s a statstcal tool for the

More information

Fourth Order Four-Stage Diagonally Implicit Runge-Kutta Method for Linear Ordinary Differential Equations ABSTRACT INTRODUCTION

Fourth Order Four-Stage Diagonally Implicit Runge-Kutta Method for Linear Ordinary Differential Equations ABSTRACT INTRODUCTION Malasa Joural of Mathematcal Sceces (): 95-05 (00) Fourth Order Four-Stage Dagoall Implct Ruge-Kutta Method for Lear Ordar Dfferetal Equatos Nur Izzat Che Jawas, Fudzah Ismal, Mohamed Sulema, 3 Azm Jaafar

More information

Generative classification models

Generative classification models CS 75 Mache Learg Lecture Geeratve classfcato models Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Data: D { d, d,.., d} d, Classfcato represets a dscrete class value Goal: lear f : X Y Bar classfcato

More information

Summary of the lecture in Biostatistics

Summary of the lecture in Biostatistics Summary of the lecture Bostatstcs Probablty Desty Fucto For a cotuos radom varable, a probablty desty fucto s a fucto such that: 0 dx a b) b a dx A probablty desty fucto provdes a smple descrpto of the

More information

Lecture Notes 2. The ability to manipulate matrices is critical in economics.

Lecture Notes 2. The ability to manipulate matrices is critical in economics. Lecture Notes. Revew of Matrces he ablt to mapulate matrces s crtcal ecoomcs.. Matr a rectagular arra of umbers, parameters, or varables placed rows ad colums. Matrces are assocated wth lear equatos. lemets

More information

CSE 5526: Introduction to Neural Networks Linear Regression

CSE 5526: Introduction to Neural Networks Linear Regression CSE 556: Itroducto to Neural Netorks Lear Regresso Part II 1 Problem statemet Part II Problem statemet Part II 3 Lear regresso th oe varable Gve a set of N pars of data , appromate d by a lear fucto

More information

Takagi-Sugeno (T-S) Fuzzy Regression of Fuzzy Data

Takagi-Sugeno (T-S) Fuzzy Regression of Fuzzy Data Proceedgs of the World Cogress o Egeerg ad Compter Scece 3 Vol I WCECS 3, 3-5 October, 3, Sa Fracsco, USA Taag-Sgeo (T-S) Fzzy Regresso of Fzzy Data Pehog Wag, Yfa Wag, ad Zhgag S * Abstract The covetoal

More information

N-dimensional Auto-Bäcklund Transformation and Exact Solutions to n-dimensional Burgers System

N-dimensional Auto-Bäcklund Transformation and Exact Solutions to n-dimensional Burgers System N-dmesoal Ato-Bäckld Trasformato ad Eact Soltos to -dmesoal Brgers System Mglag Wag Jlag Zhag * & Xagzheg L. School of Mathematcs & Statstcs Hea Uversty of Scece & Techology Loyag 4703 PR Cha. School of

More information

Generalized Linear Models. Statistical Models. Classical Linear Regression Why easy formulation if complicated formulation exists?

Generalized Linear Models. Statistical Models. Classical Linear Regression Why easy formulation if complicated formulation exists? Statstcal Models Geeralzed Lear Models Classcal lear regresso complcated formlato of smple model, strctral ad radom compoet of the model Lectre 5 Geeralzed Lear Models Geeralzed lear models geeral descrpto

More information

Analysis of Lagrange Interpolation Formula

Analysis of Lagrange Interpolation Formula P IJISET - Iteratoal Joural of Iovatve Scece, Egeerg & Techology, Vol. Issue, December 4. www.jset.com ISS 348 7968 Aalyss of Lagrage Iterpolato Formula Vjay Dahya PDepartmet of MathematcsMaharaja Surajmal

More information

Sampling Theory MODULE V LECTURE - 14 RATIO AND PRODUCT METHODS OF ESTIMATION

Sampling Theory MODULE V LECTURE - 14 RATIO AND PRODUCT METHODS OF ESTIMATION Samplg Theor MODULE V LECTUE - 4 ATIO AND PODUCT METHODS OF ESTIMATION D. SHALABH DEPATMENT OF MATHEMATICS AND STATISTICS INDIAN INSTITUTE OF TECHNOLOG KANPU A mportat objectve a statstcal estmato procedure

More information

Simple Linear Regression

Simple Linear Regression Statstcal Methods I (EST 75) Page 139 Smple Lear Regresso Smple regresso applcatos are used to ft a model descrbg a lear relatoshp betwee two varables. The aspects of least squares regresso ad correlato

More information

Correlation and Regression Analysis

Correlation and Regression Analysis Chapter V Correlato ad Regresso Aalss R. 5.. So far we have cosdered ol uvarate dstrbutos. Ma a tme, however, we come across problems whch volve two or more varables. Ths wll be the subject matter of the

More information

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin Learzato of the Swg Equato We wll cover sectos.5.-.6 ad begg of Secto 3.3 these otes. 1. Sgle mache-fte bus case Cosder a sgle mache coected to a fte bus, as show Fg. 1 below. E y1 V=1./_ Fg. 1 The admttace

More information

NumericalSimulationofWaveEquation

NumericalSimulationofWaveEquation Global Joral of Scece Froter Research: A Physcs ad Space Scece Volme 4 Isse 7 Verso. Year 4 Type : Doble Bld Peer Revewed Iteratoal Research Joral Pblsher: Global Jorals Ic. (USA Ole ISSN: 49-466 & Prt

More information

Lecture 07: Poles and Zeros

Lecture 07: Poles and Zeros Lecture 07: Poles ad Zeros Defto of poles ad zeros The trasfer fucto provdes a bass for determg mportat system respose characterstcs wthout solvg the complete dfferetal equato. As defed, the trasfer fucto

More information

An Introduction to. Support Vector Machine

An Introduction to. Support Vector Machine A Itroducto to Support Vector Mache Support Vector Mache (SVM) A classfer derved from statstcal learg theory by Vapk, et al. 99 SVM became famous whe, usg mages as put, t gave accuracy comparable to eural-etwork

More information

Discrete Adomian Decomposition Method for. Solving Burger s-huxley Equation

Discrete Adomian Decomposition Method for. Solving Burger s-huxley Equation It. J. Cotemp. Math. Sceces, Vol. 8, 03, o. 3, 63-63 HIKARI Ltd, www.m-har.com http://dx.do.org/0.988/jcms.03.3570 Dscrete Adoma Decomposto Method for Solvg Brger s-hxley Eqato Abdlghafor M. Al-Rozbaya

More information

Chapter Business Statistics: A First Course Fifth Edition. Learning Objectives. Correlation vs. Regression. In this chapter, you learn:

Chapter Business Statistics: A First Course Fifth Edition. Learning Objectives. Correlation vs. Regression. In this chapter, you learn: Chapter 3 3- Busess Statstcs: A Frst Course Ffth Edto Chapter 2 Correlato ad Smple Lear Regresso Busess Statstcs: A Frst Course, 5e 29 Pretce-Hall, Ic. Chap 2- Learg Objectves I ths chapter, you lear:

More information

Study of Correlation using Bayes Approach under bivariate Distributions

Study of Correlation using Bayes Approach under bivariate Distributions Iteratoal Joural of Scece Egeerg ad Techolog Research IJSETR Volume Issue Februar 4 Stud of Correlato usg Baes Approach uder bvarate Dstrbutos N.S.Padharkar* ad. M.N.Deshpade** *Govt.Vdarbha Isttute of

More information

Line Fitting and Regression

Line Fitting and Regression Marquette Uverst MSCS6 Le Fttg ad Regresso Dael B. Rowe, Ph.D. Professor Departmet of Mathematcs, Statstcs, ad Computer Scece Coprght 8 b Marquette Uverst Least Squares Regresso MSCS6 For LSR we have pots

More information

Isomorphism on Intuitionistic Fuzzy Directed Hypergraphs

Isomorphism on Intuitionistic Fuzzy Directed Hypergraphs Iteratoal Joral of Scetfc ad Research Pblcatos, Volme, Isse, March 0 ISSN 50-5 Isomorphsm o Ittostc Fzzy Drected Hypergraphs R.Parath*, S.Thlagaath*,K.T.Ataasso** * Departmet of Mathematcs, Vellalar College

More information

L5 Polynomial / Spline Curves

L5 Polynomial / Spline Curves L5 Polyomal / Sple Curves Cotets Coc sectos Polyomal Curves Hermte Curves Bezer Curves B-Sples No-Uform Ratoal B-Sples (NURBS) Mapulato ad Represetato of Curves Types of Curve Equatos Implct: Descrbe a

More information

Research on SVM Prediction Model Based on Chaos Theory

Research on SVM Prediction Model Based on Chaos Theory Advaced Scece ad Techology Letters Vol.3 (SoftTech 06, pp.59-63 http://dx.do.org/0.457/astl.06.3.3 Research o SVM Predcto Model Based o Chaos Theory Sog Lagog, Wu Hux, Zhag Zezhog 3, College of Iformato

More information

Fuzzy Cluster Centers Separation Clustering Using Possibilistic Approach

Fuzzy Cluster Centers Separation Clustering Using Possibilistic Approach Fzzy Clster Ceters Separato Clsterg Usg Possblstc Approach Xaohog W,,, B W 3, J S, Haj F, ad Jewe Zhao School of Food ad Bologcal Egeerg, Jags Uversty, Zhejag 03, P.R. Cha el.: +86 5887945 wxh49@js.ed.c

More information

Unimodality Tests for Global Optimization of Single Variable Functions Using Statistical Methods

Unimodality Tests for Global Optimization of Single Variable Functions Using Statistical Methods Malaysa Umodalty Joural Tests of Mathematcal for Global Optmzato Sceces (): of 05 Sgle - 5 Varable (007) Fuctos Usg Statstcal Methods Umodalty Tests for Global Optmzato of Sgle Varable Fuctos Usg Statstcal

More information

9.1 Introduction to the probit and logit models

9.1 Introduction to the probit and logit models EC3000 Ecoometrcs Lecture 9 Probt & Logt Aalss 9. Itroducto to the probt ad logt models 9. The logt model 9.3 The probt model Appedx 9. Itroducto to the probt ad logt models These models are used regressos

More information

NARMA-L2 Control of a Nonlinear Half-Car Servo-Hydraulic Vehicle Suspension System

NARMA-L2 Control of a Nonlinear Half-Car Servo-Hydraulic Vehicle Suspension System Acta Poltechca Hgarca Vol. 10 No. 4 2013 NARMA-L2 Cotrol of a Nolear Half-Car Servo-Hdralc Vehcle Sspeso Sstem Jmoh Pedro Joh Eor School of Mechacal Aeroatcal ad Idstral Egeerg Uverst of the Wtwatersrad

More information

The Finite Volume Method for Solving Systems. of Non-linear Initial-Boundary. Value Problems for PDE's

The Finite Volume Method for Solving Systems. of Non-linear Initial-Boundary. Value Problems for PDE's Appled Matematcal Sceces, Vol. 7, 13, o. 35, 1737-1755 HIKARI Ltd, www.m-ar.com Te Fte Volme Metod for Solvg Systems of No-lear Ital-Bodary Vale Problems for PDE's 1 Ema Al Hssa ad Zaab Moammed Alwa 1

More information

MULTIDIMENSIONAL HETEROGENEOUS VARIABLE PREDICTION BASED ON EXPERTS STATEMENTS. Gennadiy Lbov, Maxim Gerasimov

MULTIDIMENSIONAL HETEROGENEOUS VARIABLE PREDICTION BASED ON EXPERTS STATEMENTS. Gennadiy Lbov, Maxim Gerasimov Iteratoal Boo Seres "Iformato Scece ad Computg" 97 MULTIIMNSIONAL HTROGNOUS VARIABL PRICTION BAS ON PRTS STATMNTS Geady Lbov Maxm Gerasmov Abstract: I the wors [ ] we proposed a approach of formg a cosesus

More information

ESS Line Fitting

ESS Line Fitting ESS 5 014 17. Le Fttg A very commo problem data aalyss s lookg for relatoshpetwee dfferet parameters ad fttg les or surfaces to data. The smplest example s fttg a straght le ad we wll dscuss that here

More information

ANALYSIS ON THE NATURE OF THE BASIC EQUATIONS IN SYNERGETIC INTER-REPRESENTATION NETWORK

ANALYSIS ON THE NATURE OF THE BASIC EQUATIONS IN SYNERGETIC INTER-REPRESENTATION NETWORK Far East Joural of Appled Mathematcs Volume, Number, 2008, Pages Ths paper s avalable ole at http://www.pphm.com 2008 Pushpa Publshg House ANALYSIS ON THE NATURE OF THE ASI EQUATIONS IN SYNERGETI INTER-REPRESENTATION

More information

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions Iteratoal Joural of Computatoal Egeerg Research Vol, 0 Issue, Estmato of Stress- Stregth Relablty model usg fte mxture of expoetal dstrbutos K.Sadhya, T.S.Umamaheswar Departmet of Mathematcs, Lal Bhadur

More information

ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK

ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK Ram Rzayev Cyberetc Isttute of the Natoal Scece Academy of Azerbaa Republc ramrza@yahoo.com Aygu Alasgarova Khazar

More information

Econometric Methods. Review of Estimation

Econometric Methods. Review of Estimation Ecoometrc Methods Revew of Estmato Estmatg the populato mea Radom samplg Pot ad terval estmators Lear estmators Ubased estmators Lear Ubased Estmators (LUEs) Effcecy (mmum varace) ad Best Lear Ubased Estmators

More information

Lecture 5: Interpolation. Polynomial interpolation Rational approximation

Lecture 5: Interpolation. Polynomial interpolation Rational approximation Lecture 5: Iterpolato olyomal terpolato Ratoal appromato Coeffcets of the polyomal Iterpolato: Sometme we kow the values of a fucto f for a fte set of pots. Yet we wat to evaluate f for other values perhaps

More information

Chapter 14 Logistic Regression Models

Chapter 14 Logistic Regression Models Chapter 4 Logstc Regresso Models I the lear regresso model X β + ε, there are two types of varables explaatory varables X, X,, X k ad study varable y These varables ca be measured o a cotuous scale as

More information

CHANNEL IMPAIRMENTS & EQUALIZATION. Prepared by Deepa.T, Asst.Prof. /TCE

CHANNEL IMPAIRMENTS & EQUALIZATION. Prepared by Deepa.T, Asst.Prof. /TCE CHANNEL IMPAIRMENTS & EQUALIZATION Prepared by Deepa.T, Asst.Prof. /TCE Revew of Relevat Cocepts Fadg: 1) Flat Fadg 2) Freqecy Selectve Fadg 3) Other Mlt path Cocers Flat Fadg Flat Fadg s cased by absorbers

More information

International Journal of Scientific & Engineering Research, Volume 5, Issue 9, September ISSN

International Journal of Scientific & Engineering Research, Volume 5, Issue 9, September ISSN Iteratoal Joral o Scetc & Egeerg Research, Volme 5, Isse 9, September-4 5 ISSN 9-558 Nmercal Implemetato o BD va Method o Les or Tme Depedet Nolear Brgers Eqato VjthaMkda, Ashsh Awasth Departmet o Mathematcs,

More information

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses Johs Hopks Uverst Departmet of Bostatstcs Math Revew for Itroductor Courses Ratoale Bostatstcs courses wll rel o some fudametal mathematcal relatoshps, fuctos ad otato. The purpose of ths Math Revew s

More information

Bayes (Naïve or not) Classifiers: Generative Approach

Bayes (Naïve or not) Classifiers: Generative Approach Logstc regresso Bayes (Naïve or ot) Classfers: Geeratve Approach What do we mea by Geeratve approach: Lear p(y), p(x y) ad the apply bayes rule to compute p(y x) for makg predctos Ths s essetally makg

More information

Some Different Perspectives on Linear Least Squares

Some Different Perspectives on Linear Least Squares Soe Dfferet Perspectves o Lear Least Squares A stadard proble statstcs s to easure a respose or depedet varable, y, at fed values of oe or ore depedet varables. Soetes there ests a deterstc odel y f (,,

More information

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses Johs Hopks Uverst Departmet of Bostatstcs Math Revew for Itroductor Courses Ratoale Bostatstcs courses wll rel o some fudametal mathematcal relatoshps, fuctos ad otato. The purpose of ths Math Revew s

More information

Lecture 16: Backpropogation Algorithm Neural Networks with smooth activation functions

Lecture 16: Backpropogation Algorithm Neural Networks with smooth activation functions CO-511: Learg Theory prg 2017 Lecturer: Ro Lv Lecture 16: Bacpropogato Algorthm Dsclamer: These otes have ot bee subected to the usual scruty reserved for formal publcatos. They may be dstrbuted outsde

More information

Beam Warming Second-Order Upwind Method

Beam Warming Second-Order Upwind Method Beam Warmg Secod-Order Upwd Method Petr Valeta Jauary 6, 015 Ths documet s a part of the assessmet work for the subject 1DRP Dfferetal Equatos o Computer lectured o FNSPE CTU Prague. Abstract Ths documet

More information

On the Modeling and Simulation of Collision and Collision-Free Motion for Planar Robotic Arm Galia V. Tzvetkova

On the Modeling and Simulation of Collision and Collision-Free Motion for Planar Robotic Arm Galia V. Tzvetkova Iteratoal Joural of Egeerg Research & Scece (IJOER [Vol-, Issue-9, December- 25] O the Modelg ad Smulato of Collso ad Collso-Free Moto for Plaar Robotc Arm Gala V. Tzvetova Isttute of mechacs, Bulgara

More information

Statistics. Correlational. Dr. Ayman Eldeib. Simple Linear Regression and Correlation. SBE 304: Linear Regression & Correlation 1/3/2018

Statistics. Correlational. Dr. Ayman Eldeib. Simple Linear Regression and Correlation. SBE 304: Linear Regression & Correlation 1/3/2018 /3/08 Sstems & Bomedcal Egeerg Departmet SBE 304: Bo-Statstcs Smple Lear Regresso ad Correlato Dr. Ama Eldeb Fall 07 Descrptve Orgasg, summarsg & descrbg data Statstcs Correlatoal Relatoshps Iferetal Geeralsg

More information

Transforms that are commonly used are separable

Transforms that are commonly used are separable Trasforms s Trasforms that are commoly used are separable Eamples: Two-dmesoal DFT DCT DST adamard We ca the use -D trasforms computg the D separable trasforms: Take -D trasform of the rows > rows ( )

More information

u(x, t) = u 0 (x ct). This Riemann invariant u is constant along characteristics λ with x = x 0 +ct (u(x, t) = u 0 (x 0 )):

u(x, t) = u 0 (x ct). This Riemann invariant u is constant along characteristics λ with x = x 0 +ct (u(x, t) = u 0 (x 0 )): x, t, h x The Frst-Order Wave Eqato The frst-order wave advecto eqato s c > 0 t + c x = 0, x, t = 0 = 0x. The solto propagates the tal data 0 to the rght wth speed c: x, t = 0 x ct. Ths Rema varat s costat

More information

Fundamentals of Regression Analysis

Fundamentals of Regression Analysis Fdametals of Regresso Aalyss Regresso aalyss s cocered wth the stdy of the depedece of oe varable, the depedet varable, o oe or more other varables, the explaatory varables, wth a vew of estmatg ad/or

More information

LINEARLY CONSTRAINED MINIMIZATION BY USING NEWTON S METHOD

LINEARLY CONSTRAINED MINIMIZATION BY USING NEWTON S METHOD Jural Karya Asl Loreka Ahl Matematk Vol 8 o 205 Page 084-088 Jural Karya Asl Loreka Ahl Matematk LIEARLY COSTRAIED MIIMIZATIO BY USIG EWTO S METHOD Yosza B Dasrl, a Ismal B Moh 2 Faculty Electrocs a Computer

More information

PGE 310: Formulation and Solution in Geosystems Engineering. Dr. Balhoff. Interpolation

PGE 310: Formulation and Solution in Geosystems Engineering. Dr. Balhoff. Interpolation PGE 30: Formulato ad Soluto Geosystems Egeerg Dr. Balhoff Iterpolato Numercal Methods wth MATLAB, Recktewald, Chapter 0 ad Numercal Methods for Egeers, Chapra ad Caale, 5 th Ed., Part Fve, Chapter 8 ad

More information

Midterm Exam 1, section 1 (Solution) Thursday, February hour, 15 minutes

Midterm Exam 1, section 1 (Solution) Thursday, February hour, 15 minutes coometrcs, CON Sa Fracsco State Uversty Mchael Bar Sprg 5 Mdterm am, secto Soluto Thursday, February 6 hour, 5 mutes Name: Istructos. Ths s closed book, closed otes eam.. No calculators of ay kd are allowed..

More information

STK4011 and STK9011 Autumn 2016

STK4011 and STK9011 Autumn 2016 STK4 ad STK9 Autum 6 Pot estmato Covers (most of the followg materal from chapter 7: Secto 7.: pages 3-3 Secto 7..: pages 3-33 Secto 7..: pages 35-3 Secto 7..3: pages 34-35 Secto 7.3.: pages 33-33 Secto

More information

Inverse Problem of Finding an Unknown Parameter for One- and Two-dimensional Parabolic Heat Equations

Inverse Problem of Finding an Unknown Parameter for One- and Two-dimensional Parabolic Heat Equations Iverse Problem of Fdg a Ukow Parameter for Oe- ad Two-dmesoal Parabolc Heat Eqatos Mohamed Elmadob Problem Report sbmtted to the Statler College of Egeerg ad Meral Resorces at West Vrga Uversty partal

More information

Lecture 3. Sampling, sampling distributions, and parameter estimation

Lecture 3. Sampling, sampling distributions, and parameter estimation Lecture 3 Samplg, samplg dstrbutos, ad parameter estmato Samplg Defto Populato s defed as the collecto of all the possble observatos of terest. The collecto of observatos we take from the populato s called

More information

C.11 Bang-bang Control

C.11 Bang-bang Control Itroucto to Cotrol heory Iclug Optmal Cotrol Nguye a e -.5 C. Bag-bag Cotrol. Itroucto hs chapter eals wth the cotrol wth restrctos: s boue a mght well be possble to have scotutes. o llustrate some of

More information

0/1 INTEGER PROGRAMMING AND SEMIDEFINTE PROGRAMMING

0/1 INTEGER PROGRAMMING AND SEMIDEFINTE PROGRAMMING CONVEX OPIMIZAION AND INERIOR POIN MEHODS FINAL PROJEC / INEGER PROGRAMMING AND SEMIDEFINE PROGRAMMING b Luca Buch ad Natala Vktorova CONENS:.Itroducto.Formulato.Applcato to Kapsack Problem 4.Cuttg Plaes

More information

Analysis of Variance with Weibull Data

Analysis of Variance with Weibull Data Aalyss of Varace wth Webull Data Lahaa Watthaacheewaul Abstract I statstcal data aalyss by aalyss of varace, the usual basc assumptos are that the model s addtve ad the errors are radomly, depedetly, ad

More information

Meromorphic Solutions of Nonlinear Difference Equations

Meromorphic Solutions of Nonlinear Difference Equations Mathematcal Comptato Je 014 Volme 3 Isse PP.49-54 Meromorphc Soltos of Nolear Dfferece Eatos Xogyg L # Bh Wag College of Ecoomcs Ja Uversty Gagzho Gagdog 51063 P.R.Cha #Emal: lxogyg818@163.com Abstract

More information

Lower and upper bound for parametric Useful R-norm information measure

Lower and upper bound for parametric Useful R-norm information measure Iteratoal Joral of Statstcs ad Aled Mathematcs 206; (3): 6-20 ISS: 2456-452 Maths 206; (3): 6-20 206 Stats & Maths wwwmathsjoralcom eceved: 04-07-206 Acceted: 05-08-206 haesh Garg Satsh Kmar ower ad er

More information

Unsupervised Learning and Other Neural Networks

Unsupervised Learning and Other Neural Networks CSE 53 Soft Computg NOT PART OF THE FINAL Usupervsed Learg ad Other Neural Networs Itroducto Mture Destes ad Idetfablty ML Estmates Applcato to Normal Mtures Other Neural Networs Itroducto Prevously, all

More information

Midterm Exam 1, section 2 (Solution) Thursday, February hour, 15 minutes

Midterm Exam 1, section 2 (Solution) Thursday, February hour, 15 minutes coometrcs, CON Sa Fracsco State Uverst Mchael Bar Sprg 5 Mdterm xam, secto Soluto Thursda, Februar 6 hour, 5 mutes Name: Istructos. Ths s closed book, closed otes exam.. No calculators of a kd are allowed..

More information

Maximum Likelihood Estimation

Maximum Likelihood Estimation Marquette Uverst Maxmum Lkelhood Estmato Dael B. Rowe, Ph.D. Professor Departmet of Mathematcs, Statstcs, ad Computer Scece Coprght 08 b Marquette Uverst Maxmum Lkelhood Estmato We have bee sag that ~

More information

Lecture Notes Forecasting the process of estimating or predicting unknown situations

Lecture Notes Forecasting the process of estimating or predicting unknown situations Lecture Notes. Ecoomc Forecastg. Forecastg the process of estmatg or predctg ukow stuatos Eample usuall ecoomsts predct future ecoomc varables Forecastg apples to a varet of data () tme seres data predctg

More information

Previous lecture. Lecture 8. Learning outcomes of this lecture. Today. Statistical test and Scales of measurement. Correlation

Previous lecture. Lecture 8. Learning outcomes of this lecture. Today. Statistical test and Scales of measurement. Correlation Lecture 8 Emprcal Research Methods I434 Quattatve Data aalss II Relatos Prevous lecture Idea behd hpothess testg Is the dfferece betwee two samples a reflecto of the dfferece of two dfferet populatos or

More information

13. Parametric and Non-Parametric Uncertainties, Radial Basis Functions and Neural Network Approximations

13. Parametric and Non-Parametric Uncertainties, Radial Basis Functions and Neural Network Approximations Lecture 7 3. Parametrc ad No-Parametrc Ucertates, Radal Bass Fuctos ad Neural Network Approxmatos he parameter estmato algorthms descrbed prevous sectos were based o the assumpto that the system ucertates

More information

Open and Closed Networks of M/M/m Type Queues (Jackson s Theorem for Open and Closed Networks) Copyright 2015, Sanjay K. Bose 1

Open and Closed Networks of M/M/m Type Queues (Jackson s Theorem for Open and Closed Networks) Copyright 2015, Sanjay K. Bose 1 Ope ad Closed Networks of //m Type Qees Jackso s Theorem for Ope ad Closed Networks Copyrght 05, Saay. Bose p osso Rate λp osso rocess Average Rate λ p osso Rate λp N p p N osso Rate λp N Splttg a osso

More information

Consumer theory. A. The preference ordering B. The feasible set C. The consumption decision. A. The preference ordering. Consumption bundle

Consumer theory. A. The preference ordering B. The feasible set C. The consumption decision. A. The preference ordering. Consumption bundle Föreläsgsderlag för Gravelle-Rees. Del. Thomas Soesso Cosmer theory A. The referece orderg B. The feasble set C. The cosmto decso A. The referece orderg Cosmto bdle ( 2,,... ) Assmtos: Comleteess 2 Trastvty

More information

Introduction to local (nonparametric) density estimation. methods

Introduction to local (nonparametric) density estimation. methods Itroducto to local (oparametrc) desty estmato methods A slecture by Yu Lu for ECE 66 Sprg 014 1. Itroducto Ths slecture troduces two local desty estmato methods whch are Parze desty estmato ad k-earest

More information

Chapter Newton-Raphson Method of Solving Simultaneous Nonlinear Equations

Chapter Newton-Raphson Method of Solving Simultaneous Nonlinear Equations Chapter 7 Newto-Rapho Method o Solg Smltaeo Nolear Eqato Ater readg th chapter o hold be able to: dere the Newto-Rapho method ormla or mltaeo olear eqato deelop the algorthm o the Newto-Rapho method or

More information

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions.

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions. Ordary Least Squares egresso. Smple egresso. Algebra ad Assumptos. I ths part of the course we are gog to study a techque for aalysg the lear relatoshp betwee two varables Y ad X. We have pars of observatos

More information