A new terrain mapping method for mobile robots obstacle negotiation
|
|
- Ashlynn O’Connor’
- 6 years ago
- Views:
Transcription
1 A ew errai mappig mehod for mobile robos obsacle egoiaio Cag Ye ad Joha Boresei Advaced Techologies Lab The Uiversiy of Michiga 0 Beal Ave, A Arbor, MI, USA ABSTRACT This paper iroduces a ew errai mappig mehod for mobile robos wih a -D laser ragefider. I he proposed mehod, a elevaio map ad a ceraiy map are buil ad used for he filerig of erroeous daa. The filer, called Ceraiy Assised Spaial (CAS) filer, firs employs he physical cosrais o moio coiuiy ad spaial coiuiy o disiguish corruped pixels (e.g., due o arifacs, radom oise, or he mixed pixels effec) ad missig daa from ucorruped pixels i a elevaio map. I he removes he corruped pixels ad missig daa, while missig daa is filled i by a Weighed Media filer. Ucorruped pixels are lef iac so as o reai edges of objecs. Our exesive idoor ad oudoor mappig experimes demosrae ha he CAS filer has beer performace i erroeous daa reducio ad map deail preservaio ha exisig filers. Keywords: laser ragefider, errai mappig, filerig, elevaio map, ceraiy map, mixed pixels.. INTRODUCTION To auoomously avigae o rugged errai (i.e., idoor eviromes wih debris o he floor or oudoor, off-road eviromes), a mobile robo requires he capabiliy o decide wheher a obsacle should be raversed or circumavigaed. The abiliy o make his decisio ad o apply he appropriae acio is called obsacle egoiaio (ON). A crucial issue ivolved i ON is errai mappig. Research effors o errai mappig have bee devoed o idoor eviromes [], oudoor, off-road errai [, 3, 4], as well as plaeary errai [5, 6, 7]. Mos of he exisig mehods employ sereovisio [, 3, 4, 7], which is sesiive o eviromeal codiio (e.g., ambie illumiaio) ad has low rage resoluio ad accuracy. As a aleraive or suppleme, 3-D Laser Ragefiders (LRFs) have bee employed sice he early ieies [5, 6, 8]. However, 3-D LRFs are very cosly, bulky, ad heavy. Therefore, hey are o suiable for small ad/or expedable robos. Furhermore, mos of hem are desiged for saioary use due o he slow sca speed i elevaio. A more feasible soluio for lower-cos robos is a -D LRF. Researchers a CMU [] used a Sick -D LRF lookig upward o perform idoor 3-D mappig. Due o he upward lookig maer, he 3D map is o suiable for ON sice he groud map is o available i his case. Herick ad Krokov [9] employed a Acuiy -D LRF o compleme sereovisio for obsacle ad hazard deecio. The scaer was poied o he groud a 45. Due o he small lookahead disace, he scaer was used for safey purposes oly, i.e., whe a obsacle is deeced, a emergecy commad is issued o sop he robo. More rece research a CMU [3] also used a -D LRF o compleme sereovisio i a errai mappig applicaio. I his work a Sick LMS 0 was moued o CMU s Nomad robo so as o look diagoally dowward ad forward. As he robo moves forward, he faig laser beam swep he groud ahead of he robo ad produced rage daa of he errai. Based o his daa, he Nomad produced a so called goodess map, which was creaed based o he curre sca daa ad he previous groud level (i.e., a leas square fi lie i he previous sca). The goodess map was he combied wih he map buil by sereovisio for pah plaig. I his work, he overall map was produced maily by sereovisio. This paper preses a ew errai mappig mehod usig a sigle sesor modaliy a -D LRF, ad a iovaive filerig mehod for map buildig. The paper is orgaized as follows: I Secio we iroduce our approach o errai mappig ad we discuss sources of map misrepreseaios. We also ivesigae he problems i applyig coveioal image filers o our errai-mappig problem ad describe he priciple of our filer. I Secio 3 we prese experimeal bechmark ess performed uder corolled codiios, followed by resuls of idoor ad oudoor experimes.
2 Fig.. Experimeal seup for errai mappig sudy: he liear moio able moves he LRF a velociy v (maximum: m/s) ad he roary able ils he LRF a a maximum rae of 37.5 /s i a rage of [-5, 5 ] for roll ad [-40, -5 ] for pich. Addiioal experimes wih he LRF moued o a real mobile robo were also coduced ad are icluded i Secio 3.. MAP BUILDING AND FILTERING ALGORITHM. Map buildig priciple As depiced i Fig., a Sick LMS 00 is moued o our Gorilla vehicle such ha he LRF looks diagoally dowward ad forward (a a pich agle of - ). While he vehicle is i moio, he faig laser beam sweeps he errai ahead of he vehicle ad produce coiuous rage measuremes of he errai. For he experimes described i his paper we se our LRF o a agular resoluio of ad a serial commuicaio speed of 500 Kbaud. Uder hese codiios he LRF produces 8 rage readigs (for is 80 field of view) every 3.3 ms (see [0] for deails). Our algorihm assumes ha real-ime, 6-degree-of-freedom (DOF) pose iformaio is available from our so-called Propriocepive Pose Esimaio (PPE) sysem, which is described i deail i []. Wih PPE available, he rage daa ca be rasformed io world coordiaes ad regisered io a errai map. The errai map cosiss of a elevaio map ad a ceraiy map. Boh are -D gridype maps. Each cell i he elevaio map holds a value ha represes he heigh of cell, while each cell i he ceraiy map holds a value ha expresses our ceraiy i he accuracy of he correspodig cell i he elevaio map. Fig.. Sick LMS 00 o he Gorilla vehicle I order o projec raw rage daa oo he -D elevaio map we use he Euler rasformaio. The forward kiemaic rasformaio is give by: v v l d = Tl d () v v T l T where d = ( x, y, z) ad d = ( xl, yl, zl ) are he rage measuremes i he world coordiae sysem ad i he LRF coordiae sysem, respecively. T l is he forward kiemaic rasformaio marix, which rasforms he coordiae value of a laser rage measureme i he LRF coordiae sysem o he coordiae value i he world coordiae. For cociseess, we omi he derivaio of he rasformaio marix here. To verify his idea ad sudy he mappig characerisics, we use he experimeal seup show i Fig.. The LRF is moued o a roary able, which produces roll ad pich roaio, ad he roary able is moued o a liear moio able. This experimeal seup effecively emulaes may of he moio-iduced disorios ecouered by a vehicle raversig rugged errai. The coordiae sysems for he experimeal seup are depiced i Fig. 3. The world coordiae sysem is deoed x y z. The coordiae sysem aached o he roary able is deoed x b y b z b wih he x b coordiae beig aliged wih he pich roaio axis. y b ad z b are parallel o y ad z, respecively whe he pich agle is zero. Aoher coordiae sysem, deoed x l y l z l, is aached o he roary able so ha y l is aliged wih he roll axis ad x l ad z l are parallel o x b ad z b whe he roll agle is zero. The LRF coordiae sysem is deoed x s y s z s wih z s aliged wih z l ad x s ad y s i parallel wih x l ad y l, respecively. I view of he -degree sca resoluio ad he 80-degree field of view of he LRF, he LRF ca be hough of as havig 8 discree laser beams i x s o 3 y s, wih beam # aliged wih coordiae x s, beam #9 aimed wih y s, ad beam #8 aliged wih -x s. The coordiae values of o 3 i x l y l z l, o i x b y b z b ad o i x y z are (0, 0, p), (, m, q) ad (u, v, w), respecively. The forward kiemaic rasformaio is give by x y z lcφcβ + psφ + + u = lcθsβ + lsθsφcβ qsθ + mcθ psθcφ + v lsθsβ lcθsφcβ + qcθ + msθ + pcθcφ + w ()
3 where sx sads for si(x ) ad cx sads for cos(x ). β is he agle bewee he k h laser beam (which akes rage z measureme l) ad he x s axis ad is give by z b z β = ( k ) π /80 (3) l m For every wo cosecuive measuremes a ime seps ad +, he o q z s φ y maximum chage of heigh value z for he measureme a (x, y ) is l x l o 3 calculaed by y s z θ x s laser z z β z max = ygθ + δl + θ + φ (4) o l θ φ x b y b w o y where y is he displaceme of he LRF alog he y axis, δ l is he u maximum measureme error of he LRF (3 mm accordig o [0]), x v z / l, z / θ ad z / φ are parial derivaives ad ca be derived Fig. 3. Coordiae sysems for he experime from Eq.. I his paper a grid size of 5 mm by 5 mm is used for boh he elevaio map ad he ceraiy map. A cell i he elevaio map is deoed h(i, while a cell i he ceraiy map is deoed c(i, which is a posiive ieger smaller ha 56. Usig he above rasformaio he coordiae values x ad y are compued for each rage measureme a ime ad mapped o he grid idex i ad j, respecively. The he pixel c(i, i he ceraiy map is updaed as follows: + c (i, + 3 if z h z or (, ) = 0 + max c i j c = (5) c oherwise where he icreme of 3 is used o icrease he ceraiy value faser; ad he pixel h(i, i he elevaio map is updaed by + + z if z > h + h = (6) h oherwise I every 3.3-millisecod ierval he 8 rage measuremes acquired from he LRF are mapped io boh maps.. Sources of mappig errors Ideally, a errai map buil accordig o he above-described algorihm should be sufficie for avigaig a mobile robo. However, some rage daa from he Sick LMS 00 are erroeous ad will cause mappig errors. The pheomeo of mixed pixels, missig daa, arifacs ad oise are he mai sources of erroeous measuremes:. Mixed pixels occur whe a laser beam his he very edge of a objec so ha par of he beam is refleced from ha edge of he objec ad par of he beam is refleced from he backgroud behid he objec. The resulig rage measureme lies somewhere bewee he disace o he objec ad he disace o he backgroud [0]. I researchig he mixed pixels pheomeo we foud ha if he disace D bewee he edge of he foregroud objec ad he backgroud is close o he laser pulse widh, L ( L m for he Sick LRF ), he a subsaial umber of mixed pixels is geeraed. However, if D > L, he he umber of mixed pixels drops sigificaly. This is because he Sick LRF is desiged o accep as valid readigs oly reflecios semmig from he same pulse. This smar desig feaure is very effecive i rejecig mos ambie oise, alhough i does o elimiae all. Base o he commuicaio wih Sick Germay, he LRF is a sigle sho measureme sysem. I seds ou a pulse wih a widh close o m ad deecs he reflecio wih he same pulse widh. A a corer sho his deecig scheme allows he receiver receive reflecio from he edge ad he backgroud (wihi he meer pulse widh legh). Therefore, resul i mixed pixels. I is obvious ha due o he shor pulse widh, mixed pixels problem i he SICK LRF is much less proouced ha LRFs based o phase shif measureme echique. Ierally, he LRF uses he eergy of he received sigal o correc he raw rage measureme. This correcive scheme may cause addiioal rage error o some daa alhough i improves he accuracy of some oher daa. Because he received sigal level is deermied by rage, icidece agle ad surface refleciviy hece ca be prediced.
4 . Missig daa occur whe he measured rage is ivalid []. For isace, a refleced sigal may be o weak for deecio due o a large icidece agle ad/or due o low diffuse refleciviy of he reflecig objec. This may cause SNR oo low error. Daa may also be missig if a laser beam is rapped ad hus o refleced back o he LRF. This codiio resuls i a operaioal overflow error i he Sick LRF. Direc exposure o he suligh or similar ligh source may lead o dazzlig ad cause ivalid readigs a cerai agles. For a complee lisig of Sick s error codiios see []. 3. Eviromeal ierfereces, such as srog ambie ligh, ad shock durig moio [3] may poeially creae oisy rage measuremes ad hece resul i oise i he elevaio map. Fig. 4 shows he effecs of hese errors. Such errors i he errai map deeriorae he ON performace of he vehicle. Fig. 4. Map misrepreseaio due o rage errors: Lef: Mixed pixels ca creae phaom objecs behid he edges of objecs. Missig daa creaes empy pixels i he upper surfaces of objecs. Righ: Eviromeal ierfereces ca cause arifacs/oise. Ambie ligh ad/or shock from moio ca creae radom oise. I he 6-meer measurig mode of he Sick LRF, we observed ha here are a lo of high spikes (labeled arifacs ) i he ceer of his elevaio map, which was buil o a cerai sidewalk ear our lab. Close examiaio revealed ha mos of hese arifacs resuled from very small rage measuremes (less ha he 0 millimeer a he exreme lef ad righ ed of he Sick s 8-degree sca rage i his experime.). We were uable o deermie wha causes hese misreadigs, bu we suspec udeeced overflow o be he problem. However, i is easy eough o delee rage readigs ha are smaller ha he 0-millimeer miimum rage before hey are used for map buildig. The oher high spikes are he resul of radom oise i rage daa, which produces rage errors a several meers ahead of he LRF. (i.e., o likely o be mixed pixels)..3 Coveioal image filerig algorihms A elevaio map ca be rasformed io a rage image [4] if he heigh value of each pixel is reaed as iesiy. This way coveioal image processig filers [5, 6, 7] ca be applied o he elevaio map i a aemp o elimiae he above-discussed erroeous daa. However, he applicaio of coveioal image filerig echiques is o wihou drawbacks. The foremos reaso is ha mos coveioal image filerig mehods are applied ucodiioally across he eire rage image ad may hus blur he image (i.e., aler he value of acually correc pixels). Addiioal drawbacks exis, ad i he followig we discus idividual image filers ad heir applicabiliy o elevaio maps. The Frequecy Filer [7] (e.g., Averagig Filer, Gaussia Filer [6]) are implemeed by he covoluio of a ipu image wih a covoluio kerel i he spaial domai, i.e., each pixel i he filered image is replaced by he weighed sum of he pixels i he filer widow. The effec is ha oise is suppressed bu he edges of feaures i he image are blurred a he same ime. The Wieer Filer [7] is more effecive i oise reducio, bu i requires ha he power specrum of he oise be kow i advace ad ha he oise be idepede of he sigal. However, his requireme is o me i he case of our elevaio map where oise is highly correlaed wih he sigal, for all hree error sources. Specifically, (a) mixed pixels are always locaed behid objecs edges; (b) missig daa occur mosly whe he laser beam illumiaes a surface wih low diffuse refleciviy or a a large icidece agle. Furhermore, he Wieer Filer also eds o blurr he image sice i is based o a Fourier rasform, which is implemeed as a covoluio i he spaial domai.
5 Media filers are well kow for heir capabiliy of removig impulse oise ad preservig image edges, ad are repored o have good performace i removig mixed pixels i laser rage images [5]. The sadard Media Filer ad is variaios (Weighed Media Filer [8] ad Ceer Weighed Media Filer [9]) ofe exhibi blurrig whe a large filer widow is used, or isufficie oise reducio for a small filer widow. For isace, hi edges (because hi objecs are prese or objecs are oo high such ha oly he fro edges are sesed) i a elevaio map may be compleely removed by a sadard Media filer. Adapive Media Filers [0, ] maiai a beer balace bewee deail preservaio ad oise reducio, ad hece achieve beer performace. However, all of hese media-ype filers affec all he pixels i a image icludig ucorruped oes. A umber of media ype filers [, 3] i he lieraure may poeially disiguish corruped pixels from ucorruped oes ad apply filerig oly o he corruped pixels. These filers uilize local ipu image characerisics (spaial iformaio oly) o ideify corruped pixels. However, a limiaio of his approach is he poeeial misclassificaio of hi edges as corruped pixels ad he subseque removal of hese pixels. For isace, a hi pole i a rage image produces a lie of high iesiy pixels, which may be misierpreed by he filer as impulse oise ad removed. This suggess ha i is isufficie o use oly spaial iformaio for he ideificaio of corruped pixels..4 The Ceraiy Assised Spaial (CAS) Filer I his paper we propose a ovel filer for elevaio maps, called Ceraiy Assised Spaial (CAS) filer. This filer uilizes o oly he spaial iformaio coaied i he ufilered elevaio map, bu also he ceraiy iformaio coaied i he ceraiy map, as explaied i Secio.. Due o he coiuiy of moio, every wo cosecuive measuremes o he edge of a real objec saisfy + z h zmax. Accordig o Eq. 5, edge pixels are assiged coiuous ceraiy icremes ad resul i large ceraiy values i he ceraiy map. Therefore, edge pixels i he elevaio map ca be ideified easily. They are preserved ad o filerig is applied. Mixed pixels may occur somewhere bewee he edges of objecs ad he backgroud. Their locaios are deermied by he perceage of laser spos hiig he backgroud [0]. Whe he LRF is i moio, as i ecessarily is for mappig, his perceage chages all he ime. Cosequely mixed pixels keep + chagig heir x ad y coordiae all he ime. This meas ha he codiio z h zmax is o saisfied. Hece, mixed pixels usually have small ceraiy value ad are spaially isolaed i boh he elevaio map ad he ceraiy map. They ca be ideified ad removed easily. Arifacs ad radom oise have he same characerisics ad ca also be removed by he same priciple. Fially, missig daa creae empy pixels i a elevaio map ad heir ceraiy values are zero. We foud ha applyig a Weighed Media Filer o hose pixels wih zero ceraiy values ca supply (ierpolae) he missig daa. The filerig process works as follows: Le H={h(i, i M, j N } ad C={c(i, i M, j N } deoe he elevaio map ad he ceraiy map of a errai map, respecively, where M is he legh ad N is he widh of he map. A filer wih a widow size of (k+) (k+) (where k is a oegaive umber; i his paper k= for a widow size of 5 5) is defied symmerically surroudig he curre pixel h(i, H. The curre pixel h(i, is updaed by he oupu of he filer which is give by 0 if c( i, A ad g s = 0 ad g c = 0 y = y wm else if g s = ad c( i, = 0 (7) h( i, oherwise ad A = 3 (i( s / v ) + ) (8) where i( ) meas roud o he eares ieger, s is he grid size, v is he maximum liear velociy of he LRF which is m/s i his paper ad =3.3 ms is a ime sep. g s ad g c are he spaial coiuiy idex of H ad C for he pixels i he filer widow, respecively. They are expressed by 0 if σ > B h gs = (9) oherwise
6 ad i+ k, j + k 0 if c(m,) < D or σc(i, > E gc = m= i-k, = j k (0) oherwise h( m, ) where σ ad σ h c are he variace of he ormalized local map H L = i k m i + k, j k j + k max( h( m, )) c( m, ) ad C L = i k m i + k, j k j + k respecively. We use 0 A for D ad 0.04 for E. To keep max( h( m, )) i k, j + + k he compuaioal cos low we use H < 7 o approximae he codiio σ B i Eq. 9. y wm is he L m= i k, = j k h > oupu of a Weighed Media (WM) filer ceered a pixels (i,. The ipu pixels o he WM filer are obaied by applyig a filer skeleo [8] (as depiced i Fig. 5) o he elevaio map. For j- j- j j+ j+ isace, h(i, has oe copy while h(i-, j-) has wo copies i he ipu pixels o i- he WM filer. The use of his skeleo is based o he followig heurisic: Sice i- we suspec h(i, o be a missig daum, he eares eighborig pixels h(i±, i j±) are also likely o be missig daa, while hose pixels farher from h(i, have i+ a greaer likelihood o o be missig daa. Noe, however, ha his weigh assigme for he WM filer poeially removes more missig daa o he upper i+ surface of a objec. Fig. 5. Filer skeleo for he WM filer Whe he CAS filer is applied o he raw elevaio map, mixed pixels, arifacs, ad radom oise are ideified by examiig he moio coiuiy ( c cosraied by Eq. 5) ad he spaial coiuiy i he elevaio map ad he ceraiy map ( g s ad g c ( i, ). Missig daa is disiguished by ispecig he curre pixel s ceraiy value c ad he spaial coiuiy g s i he elevaio map. Mixed pixels, arifacs, ad radom oise are herefore removed ad he missig daa are ierpolaed by he WM filer. 3. EXPERIMENTAL RESULTS I his paper we prese exesive experimeal resuls. These resuls ca be grouped io wo caegories: Bechmark ess performed uder highly corolled codiios, ad mobiliy ess i which he LRF was moued o a mobile plaform. I all of he experimes, we used a. GHz AMD Ahlo processor-based PC ruig RT-Liux for he realime daa collecio. 3. Bechmark Tess Based o he experimeal seup i Fig., we desiged he followig four bechmark ess: ) The LRF raslaes oly (deoed T ); ) he LRF raslaes ad roaes wih roll oly (deoed TR ); 3) he LRF raslaes ad roaes wih pich oly (deoed TP); ad 4) he LRF raslaes ad roaes wih roll ad pich (deoed TRP). For he experimes described here we buil 6 obsacle courses as lised i Table I. The performace of our CAS Filer is compared wih ha of he Wieer filer, he Averagig Filer, he Media Filer, ad he Ceer Weighed Media (CWM) Filer. Fig. 6 shows he obsacle course ad Fig. 6. Obsacle course for bechmark ess Fig.7 depics he raw ad filered elevaio maps based o bechmark es T wih differe filers. I ca be observed ha he CAS filer preserves more origial pixels, removes mos of he mixed pixels, ad fill almos all of he missig daa. The Wieer Filer reais almos all missig daa, ad some mixed pixels bu i oly smoohes some oher mixed pixels. The Media Filer ad he CWM filer over-filer he poles while he Averagig Filer disor he map sigificaly.
7 TABLE I FEATURES OF THE 6 OBSTACLE COURSES Obsacle Parallelepiped Cylider Poles layou 3 4 X X -5, 7, 9 X -7, 9 3-7, 9 4 X X X,, 4, 6 5 X X -4, 6 6 X X X -4, 6 7 X X X -7 8 X X X -9 9 X X -9 0 X X X -5, 7, 9 X X X -5, 7, 9 Dimesios of parallelepiped objecs (all uis i mm): #: , #: , #3: , #4: ; Dimesios of poles: #-#9 have diameers of 4,, 9, 5,, 9, 6, 4.5 ad.5, respecively ad a legh of 94; Dimesio of cylider: diameer=4, heigh= 459 (a) Origial (raw) elevaio map (b) Elevaio map afer applyig he CAS Filer (c) Elevaio map afer applyig he Wieer Filer (d) Elevaio map afer applyig he Averagig Filer (e) Elevaio map afer applyig he Media Filer (f) Elevaio map afer applyig he CWM filer Fig. 7. Origial elevaio map ad he maps afer applyig he five esed filers I order o compare he filer performace i a quaiaive way, we measured by had he exac locaio of each obsacle i each oe of he obsacle course ad covered hese measuremes io a groud ruh elevaio map. We he compared he accuracy of each filered map agais he groud ruh map by compuig a overall idex of performace (PI) accordig o PI N M ( hf h ) j= i= = N M ( hr h ) j= i= ()
8 where h r (i,, h (i, ad h f (i, are he elevaio values of pixel (i, i he raw elevaio map, he groud ruh map, ad he filered map respecively. We arbirarily seleced M=80 ad J=00 (which has full coverage of all obsacle courses) o evaluae he experimes here. Accordig o Eq., a smaller PI meas beer filer performace. For each of he obsacle courses we performed four differe ess, amely, he T, TR, TP, ad he TRP es, as described i he begiig of his secio. The resuls are summarized i Fig. 8. I is evide from hese resuls ha our proposed CAS Filer cosisely ouperformed he four coveioal filers. The relaively poor performace of he Media Filer ad he CWM Filer is due o heir ihere over-filerig. O he oher had, whe roll ad pich moio was added o he raslaory moio, he CAS Filer performed oly margially beer ha he Wieer Filer (ad he Averagig Filer i some isaces). This is due o he followig observaios: () Roll ad/or pich moio poeially icreases mappig errors. We oiced he followig source of mappig errors: (a) Roll caused vibraios i he roary able i some posiios of he liear moio able. (b) Boh roll ad pich chaged occlusios ad he sesed heighs of he poses. This resuled i error of he absolue elevaio of he pose i he map. However, i realiy we use he physical heighs (94 mm) ad he physical cross secio areas of he poses o build he rue map. The chages i occlusio ad i he sesed heigh icrease he mappig error. Accordig o Eq., hese addiioal errors add he same value o boh he umeraors ad he deomiaors of all filer PIs. Therefore, i moves he CAS filer s PI close o ha of he Wieer ad Averagig filer as i is appare ha he relaive differece of wo fracios is reduced by addig he same addiioal value o boh umeraors ad deomiaors. 3.5 Wieer Average Media CWM CAS 3.5 Wieer Average Media CWM CAS PI.5 PI Obsacle cour se Obsacle cour se (a) Traslaio oly (b) Traslaio wih pich PI Wieer Average Media CWM CAS Obsacle course (c) Traslaio wih roll.5.5 PI Wieer Average Media CWM CAS Obsacle course (d) Traslaio wih roll & pich Fig. 8. Compariso of filerig performace () Roll ad pich moio may icrease he amou of empy pixels o he objecs upper surfaces because hey icrease he LRF s sweepig velociy whe hey are i he same direcio wih he liear able s moio. I case ha he displaceme of he laser spo (hiig he objecs upper surface) i oe ime ierval exceed 5 mm (he size of oe cell), a empy pixel will resul. Also, roll ad pich moio may icrease he icidece agle of he laser beam ad hece resul i missig daa which causes empy pixels. Whe he umber of empy pixels i he filer widow is above a cerai value, he CAS filer removes he curre pixel while he Wieer Filer or he Averagig Filer replace he curre pixel wih a spaial covoluio value. I he TRP es for he 5 h obsacle course (see Fig. 9(d)), he performace of he CAS Filer is slighly worse ha ha of he Wieer Filer because oe parallelepiped objec is so far away from he liear moio able i laeral direcio ha i creaes a large amou of missig daa o he op surface. As a resul he CAS Filer compleely removes par of he
9 surface. However, we oiced ha his oly happes whe a objec is laerally far from he LRF s pah of moio. I pracice, laerally disa objecs have lile sigificaces i obsacle egoiaio (ON) asks. Oe impora feaure of he CAS Filer is ha i preserves he edges of objecs i he elevaio map, ad edges are he key feaure for ON. We should also meio i his discussio ha, despie heir relaively good PI performace, he Wieer Filer ad Averagig Filer are o very suiable for ON. This is because he Wieer Filer does o remove arifacs/oise i elevaio maps (see Secio 3.) while he Averagig Filer eds o lower he heigh of edges i he elevaio maps a highly udesirable rai i ON applicaios. 3. Experimeal resuls o a mobile robo We carried ou a umber of idoor ad oudoor map buildig experimes wih he LRF o our Gorilla vehicle (see Fig. ). However, hese experimes were limied o sraigh lie moio o fla, horizoal groud. More diverse errai would require real-ime pose esimaio, a feaure o ye implemeed i our mobile robo. Fig. 9 shows he obsacle courses ad he vehicle i our idoor ad oudoor mappig experimes. From he maps depiced i Fig. 0, i is appare ha he CAS Filer ouperforms he Wieer Filer because i almos compleely removes mixed pixels ad fills mos of he missig daa o he op surfaces of objecs (see Fig. 0(b)). I coras, he Wieer Filer leaves some oiceable mixed pixels ad mos of he missig daa i he map, as show i Fig. 0(c). I he oudoor mappig experime (see Fig. ), he CAS Filer compleely removed all arifacs ad radom oise while he Wieer Filer did o remove ay. I is worhwhile o meio ha arifacs ca be removed by a hreshold filer ha removes readigs smaller ha he Sick s miimum rage of 0 mm. However we le hem pass hrough o our filer because Fig. 9. Obsacle courses for mappig: lef: idoor, righ: oudoor (a) Origial (raw) elevaio map (b) Elevaio map afer applyig he CAS Filer (c) Elevaio map afer applyig he Wieer Filer Fig. 0. Idoor elevaio maps (a) Origial (raw) elevaio map (b) Elevaio map afer applyig he CAS Filer (c) Elevaio map afer applyig he Wieer Filer Fig.. Oudoor elevaio maps
10 we wa o es our filer s effeciveess uder uexpeced oudoor facors. I is well kow i image processig ha radom oises (e.g., Gaussia oise) are used o ieioally corrup images, which are he used o es a image filer s oise suppressio performace. I our case, his does work because we eed ceraiy value of each pixel, i.e., oly rue experime ca be used o es our filer. The arifacs ad oise serve for his purpose. 4. DISCUSSION We are aware ha here is a aleraive o filerig mixed pixels. Adams [4] proposed a efficie mixed pixels removal mehod for phase shif LRFs. The algorihm is based o he deecio of discoiuiy of he received sigal s ampliude. However, due o he differece of measurig priciples (ime-of-fligh i he case of he Sick LRF), his mehod mus be modified o sui he Sick LRF. Such modificaio requires kowledge of he ieral physics ad he hardware desig of he LRF, ad ca hus be implemeed effecively oly by he maufacurer, o he user. Specifically, Adam s discoiuiy deecio algorihm requires several samples of he received sigal (for compuig he secod derivaive of he square of he sigal ampliude wih respec o he area illumiaed by he laser o he edge) durig he ime he laser beam raverses he edge. However, a user level we ca oly ge a sigle measureme o he edge. This meas i is impracical o impleme his algorihm. Oe migh argue ha if he Sick LRF se ou boh refleciviy iformaio ad rage daa coiuously, he similar mehod could be implemeed ouside of he Sick LRF. However, i his case he 500 Kbaud badwidh limiaio of he Sick LRF would be exceeded. This, i ur, would be force elimiaio of some rage daa a udesirable opio i mappig applicaios. Fially, our mehod is based o he characerisics of he erroeous daa ad moio coiuiy cosrai. I is device idepede. Therefore, i is more geeral ad suiable for he ed users of LRFs. Differig from Adams algorihm, our approach removes o oly mixed pixels bu also uexpeced oudoor arifacs ad oise. 5. CONCLUSIONS I his paper we preseed a ovel filerig algorihm for errai mappig wih a -D LRF. The errai map cosiss of a elevaio map ad a ceraiy map. The value of each pixel i he ceraiy map is deermied as a fucio of a socalled moio coiuiy cosrai. Typical erroeous daa, such as mixed pixels, arifacs, ad radom oise do saisfy he moio coiuiy cosrai ad hus resul i a low ceraiy value. Examiaio of he appropriae pixels i he ceraiy map ad examiaio of he spaial coiuiy i he elevaio map allows our CAS Filer o ideify ad remove erroeous pixels. Missig daa, such as ha caused by large icidece agles ad low diffuse reflecio, ca be ideified because he affeced pixels are associaed wih ceraiy values of zero ad wih poor spaial coiuiy. Oce ideified, a WM filer ca fill i he missig daa. Edges of objecs are always associaed wih relaively large ceraiy values. As a resul our CAS Filer leaves pixels represeig objec edges iac. Overall, our proposed CAS Filer produces cosisely beer resuls ha coveioal filer algorihms because i makes use of he addiioal iformaio available i he Ceraiy Map. ACKNOWLEDGMENT This work was fuded by he U.S. Deparme of Eergy uder Award No. DE-FG04-86NE3796 ad uder a gra from The Uiversiy of Michiga s Auomoive Research Ceer (ARC), fuded by TACOM. The auhors would like o hak Mr. Weber, Sick Germay for his kid help ad isighful kowledge o he LRF. The auhors would also like o hak James Berry for his work o he desig of he roary able. REFERENCES S. Thru, W. Burgard ad D. Fox, A real-ime algorihm for mobile robo mappig wih applicaio o muli-robo ad 3D mappig, Proc. IEEE Ieraioal Coferece o Roboics ad Auomaio, 000, pp
11 S. Begé-Brezez, e al, Ucerai map makig i aural eviromes, Proc. IEEE Ieraioal Coferece o Roboics ad Auomaio, 996, pp D. S. Aposolopoulos, e al, Techology ad field demosraio of roboic search for Aarcic meeories, Ieraioal Joural of Roboics Research, vol. 9, o., pp , K. Fregee, R. Madhava ad L. E. Parker, Icremeal muliage roboic mappig of oudoor errai, Proc. IEEE Ieraioal Coferece o Roboics ad Auomaio, 00, pp I. S. Kweo ad Takeo Kaade, High-resoluio errai map from muliple sesor daa, IEEE Trasacios o Paer Aalysis ad Machie Ielligece, vol. 4, o., pp. 78-9, E. Krokov ad R. Hoffma, Terrai mappig for a walkig plaeary rover, IEEE Trasacios o Roboics ad Auomaio, vol. 0, o. 6, pp , R. Simmos, e al, Experiece wih rover avigaio for luar-like errais, Proc. IEEE/RSJ Ieraioal Coferece o Iellige Robos ad Sysem, 995, pp C. M. Shoemaker ad J. A. Borsei, The demo III UGV program: a esbed for auoomous avigaio research, Proc. IEEE ISIS/CIRA/ISAS Joi Coferece, 998, pp Herikse L. ad Krokov E., Naural Terrai Hazard Deecio wih a Laser Ragefider. Proc. IEEE I. Cof. O Roboics ad Auomaio, 997, pp C. Ye ad J. Boresei, Characerizaio of a -D laser scaer for mobile robo obsacle egoiaio, Proc. IEEE Ieraioal Coferece o Roboics ad Auomaio, 00, pp L. Ojeda ad J. Boresei, "FLEXav: Fuzzy Logic Exper Rule-based Posiio Esimaio for Mobile Robos o Rugged Terrai," Proc. IEEE Ieraioal Coferece o Roboics ad Auomaio, 00, pp Sick Ic., LMS/LMI Telegram Lisig, versio 05.00, page A. Berma, J. Daya ad B. Friedlad, Improved EKF mehod of esimaig locaios wih sudde high jumps i he measureme oise, Joural of Iellige ad Roboics Sysems, vol. 3, o. 4, pp , P. J. Besl, Surfaces i rage image udersadig, Spriger-Verlag, M. Heber ad E. Krokov, 3D measuremes from imagig laser radars: how good are hey? Image ad Visio Compuig, vol. 0, o. 3, pp ,99. 6 C. F. Olso, Adapive-scale filerig ad feaurig deecio usig rage daa, IEEE Trasacios o Paer Aalysis ad Machie Ielligece, vol., o. 9, A. K. Jai, Fudameals of Digial Image Processig, Preice Hall, D. R. K. Browrigg, The weighed media filer, Commuicaio of he ACM, vol. 7, o. 8, pp , S.-J. Ko ad S.-J. Lee, Ceer weighed media filers ad heir applicaios o image ehaceme, IEEE Trasacios o Circuis ad Sysems, vol. 38, o. 9, pp , H.-M. Li ad A. N. Willso, Jr., Media filers wih adapive legh, IEEE Trasacios o Circuis ad Sysems, vol. 35, o. 6, 988. T. Che ad H. R. Wu, Applicaio of pariio-based media ype filers for suppressig oise i images, IEEE Trasacios o Image Processig, vol. 0, o. 6, pp , 00. H.-L. Eg ad K.-K. Ma, Noise adapive sof-swichig media filer, IEEE Trasacios o image processig, vol. 0, o., pp. 4-5, T. Che, K.-K. Ma ad L.-H. Che, Tri-sae media filer for image deoisig, IEEE Trasacios o Image Processig, vol. 8, o., pp , M. D. Adams ad P. J. Prober, The ierpreaio of phase ad iesiy daa from AMCW ligh deecio sesors for reliable ragig, The Ieraioal Joural of Roboics Research, vol. 5, No. 5, pp , 996.
A Novel Filter for Terrain Mapping With Laser Rangefinders
Published i IEEE Trasacios o Roboics ad Auomaio, Vol., No. 5, Oc. 4, pp. 93-9 A Novel Filer for Terrai Mappig Wih Laser Ragefiders Cag Ye, Member, IEEE ad Joha Borese Member, IEEE Absrac This paper iroduces
More informationIntroduction to Orthogonal Frequency Division Multiplexing (OFDM) Technique
Wireless Iformaio Trasmissio Sysem Lab. Iroducio o Orhogoal Frequecy Divisio Muliplexig (OFDM Techique Isiue of Commuicaios Egieerig Naioal Su Ya-se Uiversiy OFDM Overview OFDM Sysem Model Orhogoaliy Oulie
More informationIntroduction to Orthogonal Frequency Division Multiplexing (OFDM)
Wireless Iformaio Trasmissio Sysem Lab. Iroducio o Orhogoal Frequecy Divisio Muliplexig OFDM Isiue of Commuicaios Egieerig aioal Su Ya-se Uiversiy OFDM Overview OFDM Sysem Model Orhogoaliy Oulie Muli-carrier
More informationIntroduction to OFDM Systems
Wireless Iformaio Trasmissio Sysem Lab. Iroducio o OFDM Sysems Isiue of Commuicaios Egieerig Naioal Su Ya-se Uiversiy Oulie OFDM Overview OFDM Sysem Model Orhogoaliy Muli-carrier Equivale Implemeaio by
More informationHandout 10: Pulse-Code Modulation
ENGG 31-B: Priciples of Commuicaio Sysems Hadou 1: Pulse-Code Modulaio 17 18 Firs Term Isrucor: Wig-Ki Ma November 1, 17 Suggesed Readig: Chaper 7 of Simo Hayki ad Michael Moher, Commuicaio Sysems (5h
More informationComparison of Color Features for Image Retrieval
Compariso of Color Feaures for Image Rerieval S.R. Kodiuwakku 1 Deparme of Saisics & Compuer Sciece, Uiversiy of Peradeiya salukak@pd.ac.lk S.Selvarajah Deparme of Physical Sciece, Vavuiya Campus, Uiversiy
More informationIn Color We Live C IP E I. In Color We Live. t n. In Color We Live
I r We Live Eviro me I r We Live IP E I 0 12 I 2 T E vir ome 0 12 I 2 T IPEI Evirome IPEI 012 I 2 T I r We Live Iesiy Idepede RGB-o-XYZ ur amera alibraio Bria FUNT, ad Pouya BSTNI School of ompuig Sciece,
More informationAn Optimization of Gaussian UWB Pulses
0 h Ieraioal Coferece o DEVELOPMENT AND APPLICATION SYSTEMS, Suceava, Romaia, May 7-9, 00 A Opimizaio of Gaussia UWB Pulses Adria POPA "Gheorghe Asachi" Techical Uiversiy of Iaşi Bd.Carol No., RO0.506
More informationWalsh Function Based Reconstruction Method of. Command Waveform in PWM Inverter
Walsh Fucio Based Recosrucio Mehod of Commad Waveform i PWM Iverer Kri Choeisai ad Seiji Kodo Nagaoka Uiversiy of echology 63- Kamiomioka, Nagaoka 94-288, Japa Fa: +8-28-4-9, Phoe: +8-28-4-9 E-Mail: kri@s.agaokau.ac.jp
More informationField-Effect Transistors (FETs) 3.4 The p-channel MOSFET & COMS
Field-Effec Trasisors (FETs 3.4 The p-chael MOSFET & COMS A p-chael ehaceme-ype MOSFET is fabricaed o a -ype subsrae wih p+ regios for he drai ad source, ad has holes as charge carriers. The deice operaes
More informationExperiment 2: Waveform Analysis
Specrum aalyzer Experime : Waveform Aalysis 8/3/6 his experime is o familiarize you wih he aalysis of sigals ad eworks usig he HP 358A Specrum Aalyzer. he aalysis of sigals is achieved i he ope loop mode
More informationLimb-Vector Paralleling: A General Approach to Translate Postures from Human to Humanoid Robots
Limb-Vecor Parallelig: A Geeral Approach o Traslae Posures from Huma o Humaoid Robos Zhuo-Li Bao, Shih-Li Che, Chi-Chag Tug, Rayi Yau Tara, Hadziq Fabroyir ad Wei-Chug Teg Absrac evelopig iuiive ierface
More informationExpanding the Twin Pulse Swing Free Profile
8 IEEE Ieraioal Coferece o Roboics ad Auomaio Pasadea, CA, USA, May 19-3, 8 Expadig he Twi Swig Free Profile David Bowlig, Gregory Sarr, Joh Wood ad Ro Lumia Deparme of Mechaical Egieerig The Uiversiy
More informationEE123 Digital Signal Processing
Aoucemes EE3 Digial Sigal Processig Lecure Miderm: Friday ex week Ope everyhig... bu chea shee recommeded isead Who ca o say ill pm? Opioal homework ex week Will give you miderm ad pracice quesios How
More informationSelf-Interference Cancellation in Full-Duplex Radio Transceivers with Oscillator Phase Noise
Self-Ierferece Cacellaio i Full-Duplex Radio Trasceivers wih Oscillaor Phase Noise Ville Syrjälä ad Koji Yamamoo Graduae School of Iformaics Kyoo Uiversiy Yoshida-homachi Sakyo-ku Kyoo-shi 606-8501 Kyoo
More informationDetection of Pulsed Radar in a Time Division Duplexed System
Deecio of Pulsed Radar i a Time Divisio Duplexed Sysem Brad W. Zarikoff Hamilo Isiue Naioal Uiversiy of Irelad Mayooh Email: Brad.Zarikoff@uim.ie David Weldo Wireless R&D, Vecima Neworks, Ic. Vicoria,
More informationMULTI USER DETECTION FOR CDMA- OFDM/OQAM SYSTEM COMBINED WITH SPACE TIME CODING
Ieraioal Joural of Wireless & Mobile Neworks (IJWMN) Vol. 4, No. 4, Augus 212 MULTI USER DETECTION FOR CDMA- OFDM/OQAM SYSTEM COMBINED WITH SPACE TIME CODING Radhia GHARSALLAH 1 ad Ridha BOUALLEGUE 2 1
More informationITERATIVE RECEIVER FOR SPACE-TIME CODED FLAT FADING CHANNELS. Noura Sellami yz, Mohamed Siala y, David Declercq z, Inbar Fijalkow z
ITERATIVE RECEIVER FOR SPACE-TIME CODED FLAT FADING CHANNELS Noura Sellami yz, Mohamed Siala y, David Declercq z, Ibar Fijalkow z y Frace Telecom R&D, 38-40 rue du Gééral Leclerc 92794 Issy les Moulieaux,
More informationA New Adaptive Medical Image Coding with Orthogonal Polynomials
Ieraioal Joural of Compuer Applicaios (975 8887) Volume 46 No., May A New Adapive Medical Image Codig wih Orhogoal Polyomials Krishamoorhy R Deparme of CSE Aa Uiversiy of Techology - Tiruchirappalli Tiruchirappalli,
More informationA New Power System Oscillation Type Identification Method Based on Empirical Mode Decomposition and Hilbert Transform
Proceedigs of he World Cogress o Egieerig ad Compuer Sciece 4 Vol I WCECS 4, -4 Ocober, 4, Sa Fracisco, USA A New Power Sysem Oscillaio Type Ideificaio Mehod Based o Empirical Mode Decomposiio ad Hilber
More informationAn optimization algorithm for the calculation of the electrostatic discharge current equations parameters
WSEAS TRANSACTIONS o CIRCUITS ad SYSTEMS Vasiliki Via George P. Fois Lambros Ekoomou A opimizaio algorihm for he calculaio of he elecrosaic discharge curre equaios parameers VASILIKI VITA GEORGE P. FOTIS
More informationBézier curves with shape parameter *
Wag e al. / J Zhejiag Uiv SCI 5 6A(6):497-5 497 Joural of Zhejiag Uiversiy SCIENCE ISSN 9-395 hp://www.zju.edu.c/jzus E-mail: jzus@zju.edu.c Bézier curves wih shape parameer * WANG We-ao ( 王文涛 ) WANG Guo-zhao
More informationGeo-Encryption Protocol For Mobile Networks
Geo-Ecrypio Proocol For Mobile Neworks Ala Al-Fuqaha, Omar Al-Ibrahim, Ammar Rayes * Deparme of Compuer Sciece Weser Michiga Uiversiy Kalamazoo, MI {alfuqaha, oalibrah } @cs.wmich.edu Absrac We propose
More informationJune : 2016 (CBCS) Negative terminal for forward bias. Positive terminal for forward bias. Cathode ( n-type)
s Year : Commo o all raches Jue : 6 (CCS) Noe : Max. marks : 6 (i) (ii) (iii) (iv) emp ay five quesios ou of seve quesios. ll quesios carry equal marks. Draw he ea diagram, wheever ecessary. ssume daa,
More informationDigital Communications (by B. Sklar)
Digial Commuicaios by B. Sklar Sug Ho Cho Voice 2 222 39 Cellular 11 412 5178 E mail: drago@hayag.ac.kr Coes 1 Sigals ad Specra Formaig ad Basebad Trasmissio Badpass Modulaio ad Demodulaio Chael Codig
More informationON THE QAM PARALLEL TURBO-TCM SCHEMES USING RECURSIVE CONVOLUTIONAL GF(2 N ) ENCODERS
8h Europea Sigal Processig Coferece (EUSIPCO-00) Aalborg, Demark, Augus 3-7, 00 O THE QAM PARAE TURBO-TCM SCHEMES USIG RECURSIVE COVOUTIOA GF( ) ECODERS Adria Flori Pău, Cǎli Vlǎdeau, Io Marghescu, Safwa
More informationPower-Efficient Radio Resource Allocation for Low- Medium-Altitude Aerial Platform Based TD-LTE Networks
ower-efficie Radio Resource Allocaio for Low- Medium-Aliude Aerial laform Based TD-LTE eworks Liqiag Zhao, 2, Chi Zhag, Haili Zhag ad Xiaohui Li. Sae Key Laboraory of Iegraed Service eworks, Xidia Uiversiy,
More informationFast and Accurate Behavioral Simulation of Fractional-N Frequency Synthesizers and other PLL/DLL Circuits
Fas ad Accurae Behavioral Simulaio of Fracioal-N Frequecy Syhesizers ad oher PLL/LL Circuis Michael H. Perro Microsysems Techology Laboraory, MIT perro@mi.edu, hp://www-ml.mi.edu/ perro ABSTRACT Techiques
More informationResearch on Reliability for Wind Turbine Transmission System Based on Random Theory
Advaced ciece ad Techology Leers Vol8 (CT 5) pp3-9 hp://ddoiorg/457/asl58 Research o Reliabiliy or Wid Turbie Trasmissio ysem Based o Radom Theory Xi Gua Huadog Wag 3 Zhili u Da Zhao chool o Mechaical
More informationTransmit Power Efficiency of Multi-Hop Hybrid Selection/MRC Diversity for a DS-CDMA Virtual Cellular Network
Trasmi ower Efficiecy of Muli-Hop Hybri Selecio/ Diversiy for a DS-CDMA Virual Cellular ework Imae DAOU, Eisuke KUDOH a Fumiyuki ADACHI Dep. of Elecrical a Commuicaio Egieerig, Grauae School of Egieerig,
More informationMultiuser Precoding for UWB Sensor Networks
Muliuser Precodig for UWB Sesor Neworks Heirich Luecke ad Armi Wiebe Commuicaio Techology Laboraory, ETH Zurich, 89 Zurich, Swizerlad {lueckeh, wiebe}@ari.ee.ehz.ch Absrac We cosider he dowlik of a UlraWidebad
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 informationR. W. Erickson. Department of Electrical, Computer, and Energy Engineering University of Colorado, Boulder
R. W. Erickso Deparme of Elecrical, Compuer, ad Eergy Egieerig Uiversiy of Colorado, Boulder 4.2.2. The Power MOSFET Gae Source Gae leghs approachig oe micro p - p Cosiss of may small ehacememode parallelcoeced
More informationLab 3 Acceleration. What You Need To Know: Physics 211 Lab
b Lab 3 Acceleraion Wha You Need To Know: The Physics In he previous lab you learned ha he velociy of an objec can be deermined by finding he slope of he objec s posiion vs. ime graph. x v ave. = v ave.
More informationPerformance Analysis of an Ultra-Tightly Integrated GPS and Reduced IMU System
Performace Aalysis of a Ulra-Tighly Iegraed GPS ad Reduced IMU Sysem M.G. Peovello, D. Su, G. Lachapelle, M.E. Cao Posiio, Locaio ad Navigaio (PLAN) Group Deparme of Geomaics Egieerig Schulich School of
More informationA Channel Estimation Method for MIMO-OFDM Mobile WiMax Systems
A Chael Eimaio Mehod for MIMO-OFDM Mobile WiMax Syem Fabie Delere ad Yichuag Su School of Egieerig ad echology Uiveriy of Herfordhire, Hafield, Her, AL0 9AB, UK Abrac I hi paper, chael eimaio for Space-ime
More informationMoment-Based Bound on Peak-to-Average Power Ratio and Reduction with Unitary Matrix
Mome-Based Boud o Peak-o-Average Power Raio ad Reducio wih Uiary Marix Hirofumi Tsuda Deparme of Applied Mahemaics ad Physics Graduae School of Iformaics Kyoo Uiversiy Kyoo Japa suda.hirofumi.38u@s.kyoo-u.ac.jp
More informationImproving the energy efficiency for the WBSN bottleneck zone based on random linear network coding
Improvig he eergy efficiecy for he WBSN boleeck zoe based o radom liear ework codig Al Shahee, H ad Takruri Rizk, H hp://dx.doi.org/10.1049/ie wss.2017.0056 Tile Auhors Type URL Improvig he eergy efficiecy
More informationPhase-Shifting Control of Double Pulse in Harmonic Elimination Wei Peng1, a*, Junhong Zhang1, Jianxin gao1, b, Guangyi Li1, c
Inernaional Symposium on Mechanical Engineering and Maerial Science (ISMEMS 016 Phase-Shifing Conrol of Double Pulse in Harmonic Eliminaion Wei Peng1, a*, Junhong Zhang1, Jianxin gao1, b, Guangyi i1, c
More informationMIMO-OFDM with Pilot-Aided Channel Estimation for WiMax Systems
Fir Ieraioal Workhop o he Performace Ehaceme i MIMO-OFDM Syem MIMO-OFDM wih Pilo-Aided Chael Eimaio for WiMax Syem Fabie Delere ad Yichuag Su School of Egieerig ad echology Uiveriy of Herfordhire, Hafield,
More informationEECE 301 Signals & Systems Prof. Mark Fowler
EECE 3 Signals & Sysems Prof. Mark Fowler Noe Se #8 C-T Sysems: Frequency-Domain Analysis of Sysems Reading Assignmen: Secion 5.2 of Kamen and Heck /2 Course Flow Diagram The arrows here show concepual
More informationTECHNIQUE FOR TRACKING AND VISUALIZATION OF MOTION IN SEQUENCE OF IMAGES OF THE SOLAR CORONA
Proc. VI Serbia-ulgaria Asroomical Coferece elgrade 7-11 Ma 2008 Eds. M. S. Dimirijevi M. Tsveov L.. Popovi V. Golev Publ. Asr. Soc. "Rudjer ošovi" No. 9 2009 207-216 TECHNIQUE FOR TRACKING AND VISUALIZATION
More informationweight: amplitude of sine curve
Joseph Fourier s claim: all signals are sums of sinusoids of differen frequencies. weighed sine curves weigh: ampliude of sine curve all : no exacly bu doesn maer for us in pracice Example: 3 sin() + sin(*)
More informationTechnology Trends & Issues in High-Speed Digital Systems
Deailed comparison of dynamic range beween a vecor nework analyzer and sampling oscilloscope based ime domain reflecomeer by normalizing measuremen ime Sho Okuyama Technology Trends & Issues in High-Speed
More informationImproved Performance for Color to Gray and Back For Orthogonal transforms using Normalization
Ieraioal Joral of Compaioal Egieerig Research Vol 3 Isse 5 Improved Performace for Color o Gray ad Bac For Orhogoal rasforms sig ormalizaio Dr. H. B. Kere Dr. Sdeep D. Thepade Raesh. Charvedi 3 Sr. Prof.
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 informationSpring Localization I. Roland Siegwart, Margarita Chli, Martin Rufli. ASL Autonomous Systems Lab. Autonomous Mobile Robots
Spring 2017 Localizaion I Localizaion I 10.04.2017 1 2 ASL Auonomous Sysems Lab knowledge, daa base mission commands Localizaion Map Building environmen model local map posiion global map Cogniion Pah
More informationIEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 32, NO. 5, MAY
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 32, NO. 5, MAY 2014 1 Updae-Efficiecy ad Local Repairabiliy Limis for Capaciy Approachig Codes Arya Mazumdar Member, IEEE, Veka Chadar, ad Gregory
More informationPower Conversion and Signal Transmission Integration Method Based on Dual Modulation of DC-DC Converters
IEEE TRANSATIONS ON INDUSTRIAL ELETRONIS 1 Power oversio ad Sigal Trasmissio Iegraio Mehod Based o Dual Modulaio of D-D overers Jiade Wu, Member, IEEE, Ji Du, Sude Member, IEEE, Zhegyu Li, Seior Member,
More information5 Spatial Relations on Lines
5 Spaial Relaions on Lines There are number of useful problems ha can be solved wih he basic consrucion echniques developed hus far. We now look a cerain problems, which involve spaial relaionships beween
More informationSolution of ECE 342 Test 2 S12
Soluion of ECE 342 Tes 2 S2. All quesions regarding superheerodyne receivers refer o his diagram. x c () Anenna B T < B RF < 2 f B = B T Oher Signals f c Mixer f Baseband x RFi RF () x RFo () () () x i
More informationPointwise Image Operations
Poinwise Image Operaions Binary Image Analysis Jana Kosecka hp://cs.gmu.edu/~kosecka/cs482.hml - Lookup able mach image inensiy o he displayed brighness values Manipulaion of he lookup able differen Visual
More informationCommunications II Lecture 7: Performance of digital modulation
Communicaions II Lecure 7: Performance of digial modulaion Professor Kin K. Leung EEE and Compuing Deparmens Imperial College London Copyrigh reserved Ouline Digial modulaion and demodulaion Error probabiliy
More informationLecture 11. Digital Transmission Fundamentals
CS4/MSc Compuer Neworking Lecure 11 Digial Transmission Fundamenals Compuer Neworking, Copyrigh Universiy of Edinburgh 2005 Digial Transmission Fundamenals Neworks consruced ou of Links or ransmission
More informationECMA st Edition / June Near Field Communication Wired Interface (NFC-WI)
ECMA-373 1 s Ediion / June 2006 Near Field Communicaion Wired Inerface (NFC-WI) Sandard ECMA-373 1 s Ediion / June 2006 Near Field Communicaion Wired Inerface (NFC-WI) Ecma Inernaional Rue du Rhône 114
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 informationMEDIUM VOLTAGE CABLES TO IEC60502, BS6622, VDE0276
IEC2, BS22, VDE2 Sigle Core Cables o VDE 2 The sigle core cables are desiged for disribuio of elecrical power wih Applicaio: omial volage Uo/U ragig from./.kv o 9/KV ad frequecy Hz. They are suiable for
More informationMemorandum on Impulse Winding Tester
Memorandum on Impulse Winding Teser. Esimaion of Inducance by Impulse Response When he volage response is observed afer connecing an elecric charge sored up in he capaciy C o he coil L (including he inside
More informationEXPERIMENT #4 AM MODULATOR AND POWER AMPLIFIER
EXPERIMENT #4 AM MODULATOR AND POWER AMPLIFIER INTRODUCTION: Being able o ransmi a radio frequency carrier across space is of no use unless we can place informaion or inelligence upon i. This las ransmier
More information(This lesson plan assumes the students are using an air-powered rocket as described in the Materials section.)
The Mah Projecs Journal Page 1 PROJECT MISSION o MArs inroducion Many sae mah sandards and mos curricula involving quadraic equaions require sudens o solve "falling objec" or "projecile" problems, which
More informationEE 40 Final Project Basic Circuit
EE 0 Spring 2006 Final Projec EE 0 Final Projec Basic Circui Par I: General insrucion 1. The final projec will coun 0% of he lab grading, since i s going o ake lab sessions. All oher individual labs will
More informationTable of Contents. 3.0 SMPS Topologies. For Further Research. 3.1 Basic Components. 3.2 Buck (Step Down) 3.3 Boost (Step Up) 3.4 Inverter (Buck/Boost)
Table of Conens 3.0 SMPS Topologies 3.1 Basic Componens 3.2 Buck (Sep Down) 3.3 Boos (Sep Up) 3.4 nverer (Buck/Boos) 3.5 Flyback Converer 3.6 Curren Boosed Boos 3.7 Curren Boosed Buck 3.8 Forward Converer
More informationEE368/CS232 Digital Image Processing Winter Homework #1 Released: Monday, January 8 Due: Wednesday, January 17, 1:30pm
EE368/CS232 Digial Image Processing Winer 207-208 Lecure Review and Quizzes (Due: Wednesday, January 7, :30pm) Please review wha you have learned in class and hen complee he online quiz quesions for he
More informationModulation exercises. Chapter 3
Chaper 3 Modulaion exercises Each problem is annoaed wih he leer E, T, C which sands for exercise, requires some hough, requires some concepualizaion. Problems labeled E are usually mechanical, hose labeled
More informationMean Waiting and Turnaround Time for Multiple Web Transaction
Mea Waiig ad uraroud ie for Muliple Web rasacio Yog-Ji Lee Dep. of Copuer Sciece Woosog Uiversiy aejo, 300-78, Korea yjlee@woosog.ac.kr M. Aiquzzaa School of Copuer Sciece Uiversiy of Oklahoa Nora, OK
More information4.5 Biasing in BJT Amplifier Circuits
4/5/011 secion 4_5 Biasing in MOS Amplifier Circuis 1/ 4.5 Biasing in BJT Amplifier Circuis eading Assignmen: 8086 Now le s examine how we C bias MOSFETs amplifiers! f we don bias properly, disorion can
More informationDimensions. Model Number. Electrical connection emitter. Features. Electrical connection receiver. Product information. Indicators/operating means
OBE-R-SE Dimensions.8.8 ø..75 7.5 6. 5 6.7 4.9 4. 5.9 ø.6 Model Number OBE-R-SE Elecrical connecion emier Thru-beam sensor wih m fixed cable Feaures 45 cable oule for maximum mouning freedom under exremely
More information2600 Capitol Avenue Suite 200 Sacramento, CA phone fax
26 Capiol Avenue Suie 2 Sacrameno, CA 9816 916.64.4 phone 916.64.41 fax www.esassoc.com memorandum dae Sepember 2, 216 o from subjec Richard Rich, Ciy of Sacrameno; Jeffrey Dorso, Pioneer Law Group Brian
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 informationIN SOUTHERN AFRICA MICHAELGOCHFELD
Codor 83:243251 0 The Cooper Omihologd Sociey 1981 NEST SITE SELECTION IN SOUTHERN AFRICA BY KELP GULLS JOANNA AND BURGER MICHAELGOCHFELD ABSTRACT.We sudied six coloies of Kelp Gulls (Larus domiicaus)
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 informationNegative frequency communication
Negaive frequency communicaion Fanping DU Email: dufanping@homail.com Qing Huo Liu arxiv:2.43v5 [cs.it] 26 Sep 2 Deparmen of Elecrical and Compuer Engineering Duke Universiy Email: Qing.Liu@duke.edu Absrac
More information10. The Series Resistor and Inductor Circuit
Elecronicsab.nb 1. he Series esisor and Inducor Circui Inroducion he las laboraory involved a resisor, and capacior, C in series wih a baery swich on or off. I was simpler, as a pracical maer, o replace
More informationEE 330 Lecture 24. Amplification with Transistor Circuits Small Signal Modelling
EE 330 Lecure 24 Amplificaion wih Transisor Circuis Small Signal Modelling Review from las ime Area Comparison beween BJT and MOSFET BJT Area = 3600 l 2 n-channel MOSFET Area = 168 l 2 Area Raio = 21:1
More informationECE ANALOG COMMUNICATIONS - INVESTIGATION 7 INTRODUCTION TO AMPLITUDE MODULATION - PART II
ECE 405 - ANALOG COMMUNICATIONS - INVESTIGATION 7 INTRODUCTION TO AMPLITUDE MODULATION - PART II FALL 2005 A.P. FELZER To do "well" on his invesigaion you mus no only ge he righ answers bu mus also do
More informationResearch Article Improving the Performance of Modular Production in the Apparel Assembly: A Mathematical Programming Approach
Mahemaical Problems i Egieerig, Aricle ID 472781, 7 pages hp://dx.doi.org/1.1155/214/472781 Research Aricle Improvig he Performace of Modular Producio i he Apparel Assembly: A Mahemaical Programmig Approach
More informationPulse Train Controlled PCCM Buck-Boost Converter Ming Qina, Fangfang Lib
5h Inernaional Conference on Environmen, Maerials, Chemisry and Power Elecronics (EMCPE 016 Pulse Train Conrolled PCCM Buck-Boos Converer Ming Qina, Fangfang ib School of Elecrical Engineering, Zhengzhou
More informationAnalysis of Current Starved Voltage Controlled Oscillator using 45nm CMOS Technology
SSN (Pri) : 30 3765 SSN (Olie): 78 8875 eraioal Joural of Advaced Research i Elecrical, Elecroics ad srumeaio Egieerig (A SO 397: 007 erified Orgaizaio) ol. 3, ssue 3, March 014 Aalysis of urre Sarved
More informationMATLAB/SIMULINK TECHNOLOGY OF THE SYGNAL MODULATION
J Modern Technology & Engineering Vol2, No1, 217, pp76-81 MATLAB/SIMULINK TECHNOLOGY OF THE SYGNAL MODULATION GA Rusamov 1*, RJ Gasimov 1, VG Farhadov 1 1 Azerbaijan Technical Universiy, Baku, Azerbaijan
More informationImplementation of Variable Duty Cycle Ring Oscillator for Unique Identification
Imlemeaio of Variable Duy Cycle Rig Oscillaor for Uique Ideificaio N.Sivasakari (), B.Balambiga () ad P.Idumahi (3) () Assisa Professor, Dearme of Elecroics ad Commuicaio Egieerig Meco Schlek college of
More informationLecture #7: Discrete-time Signals and Sampling
EEL335: Discree-Time Signals and Sysems Lecure #7: Discree-ime Signals and Sampling. Inroducion Lecure #7: Discree-ime Signals and Sampling Unlike coninuous-ime signals, discree-ime signals have defined
More informationJeremiah: The Face of Computer Vision
Jeremiah: The Face of Compuer Visio Richard Bowde CVSSP, School of ECM Guildford Surrey +44 (0)1483 689838 r.bowde@surrey.ac.uk ABSTRACT This paper preses a humaoid compuer ierface (Jeremiah) ha is capable
More informationP. Bruschi: Project guidelines PSM Project guidelines.
Projec guidelines. 1. Rules for he execuion of he projecs Projecs are opional. Their aim is o improve he sudens knowledge of he basic full-cusom design flow. The final score of he exam is no affeced by
More informationDirect Analysis of Wave Digital Network of Microstrip Structure with Step Discontinuities
Direc Analysis of Wave Digial Nework of Microsrip Srucure wih Sep Disconinuiies BILJANA P. SOŠIĆ Faculy of Elecronic Engineering Universiy of Niš Aleksandra Medvedeva 4, Niš SERBIA MIODRAG V. GMIROVIĆ
More informationSignals and the frequency domain ENGR 40M lecture notes July 31, 2017 Chuan-Zheng Lee, Stanford University
Signals and he requency domain ENGR 40M lecure noes July 3, 07 Chuan-Zheng Lee, Sanord Universiy signal is a uncion, in he mahemaical sense, normally a uncion o ime. We oen reer o uncions as signals o
More informationECE3204 Microelectronics II Bitar / McNeill. ECE 3204 / Term D-2017 Problem Set 7
EE3204 Microelecronics II Biar / McNeill Due: Monday, May 1, 2017 EE 3204 / Term D-2017 Problem Se 7 All ex problems from Sedra and Smih, Microelecronic ircuis, 7h ediion. NOTES: Be sure your NAME and
More informationEXPERIMENT #9 FIBER OPTIC COMMUNICATIONS LINK
EXPERIMENT #9 FIBER OPTIC COMMUNICATIONS LINK INTRODUCTION: Much of daa communicaions is concerned wih sending digial informaion hrough sysems ha normally only pass analog signals. A elephone line is such
More informationElectrical connection
Reference scanner Dimensioned drawing en 02-2014/06 50117040-01 200 500mm Disance on background/reference 10-30 V DC We reserve he righ o make changes DS_HRTR46Bref_en_50117040_01.fm Robus objec deecion
More informationVolume Author/Editor: Simon Kuznets, assisted by Elizabeth Jenks. Volume URL:
This PDF is a selecion from an ou-of-prin volume from he Naional Bureau of Economic Research Volume Tile: Shares of Upper Income Groups in Income and Savings Volume Auhor/Edior: Simon Kuznes, assised by
More informationDimensions. Transmitter Receiver ø2.6. Electrical connection. Transmitter +UB 0 V. Emitter selection. = Light on = Dark on
OBE-R-SE Dimensions Transmier.. 7.5 9..5.8 4.9 4 5 M 8.9 7.5 9..5.8 4 5 M 8.9 ø.6 ø.6 Model Number OBE-R-SE Thru-beam sensor wih m fixed cable Elecrical connecion Transmier Feaures BN +UB WH IN Ulra-small
More informationImage Denoising using Fast Bilateral Filter
Iteratioal Joural of Curret Egieerig ad Techology E-ISS 77 4106, P-ISS 347 5161 015IPRESSCO, All Rights Reserved Available at http://ipressco.com/category/ijcet Research Article Riku A Joseph * ad Maju
More informationDimensions. Transmitter Receiver ø2.6. Electrical connection. Transmitter +UB 0 V. Emitter selection. = Light on = Dark on
OBE-R-SE Dimensions Transmier.. 7.5 9..5.8 4.9 4 5 M 8.9 7.5 9..5.8 4 5 M 8.9 ø.6 ø.6 Model Number OBE-R-SE Thru-beam sensor wih m fixed cable Elecrical connecion Transmier Feaures BN +UB WH IN Ulra-small
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 informationLecture September 6, 2011
cs294-p29 Seminar on Algorihmic Game heory Sepember 6, 2011 Lecure Sepember 6, 2011 Lecurer: Chrisos H. Papadimiriou Scribes: Aloni Cohen and James Andrews 1 Game Represenaion 1.1 abular Form and he Problem
More informationNoise Reduction/Mode Isolation with Adaptive Down Conversion (ADC)
Page 1 Noise Reducion/Mode Isolaion wih Adapive Down Conversion (ADC) Abel B. Diaz, Thomas W. Tunnell NSTec Los Alamos Operaions Presened o PDV Workshop 8-16-2007 Page 2 Summary Adapive down conversion
More informationRevision: June 11, E Main Suite D Pullman, WA (509) Voice and Fax
2.5.3: Sinusoidal Signals and Complex Exponenials Revision: June 11, 2010 215 E Main Suie D Pullman, W 99163 (509) 334 6306 Voice and Fax Overview Sinusoidal signals and complex exponenials are exremely
More informationUniversal microprocessor-based ON/OFF and P programmable controller MS8122A MS8122B
COMPETENCE IN MEASUREMENT Universal microprocessor-based ON/OFF and P programmable conroller MS8122A MS8122B TECHNICAL DESCRIPTION AND INSTRUCTION FOR USE PLOVDIV 2003 1 I. TECHNICAL DATA Analog inpus
More informationG.711.1: A WIDEBAND EXTENSION TO ITU-T G.711
16h Europea Sigal Processig Coferece (EUSIPCO 008), Lausae, Swizerlad, Augus 59, 008, copyrigh by EURASIP G.711.1: A WIDEBAND ETENSION TO ITUT G.711 Yusue Hiwasai 1, Shigeai Sasai 1, Hioshi Ohmuro 1, Taeshi
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 informationThe University of Melbourne Department of Mathematics and Statistics School Mathematics Competition, 2013 JUNIOR DIVISION Time allowed: Two hours
The Universiy of Melbourne Deparmen of Mahemaics and Saisics School Mahemaics Compeiion, 203 JUNIOR DIVISION Time allowed: Two hours These quesions are designed o es your abiliy o analyse a problem and
More information