Sensitivity Analysis Using Neural Network for Estimating Aircraft Stability and Control Derivatives
|
|
- Conrad Reed
- 5 years ago
- Views:
Transcription
1 Internatona Conference on Integent and Advanced Systems 27 Senstvty Anayss Usng Neura Networ for Estmatng Arcraft Stabty and Contro Dervatves Roht Garhwa a, Abhshe Hader b and Dr. Manoranan Snha c Department of Aerospace Engneerng, Indan Insttute of Technoogy, Kharagpur 72 32, INDIA e-ma: a ur.roht23@gma.com; b hader.abhshe@gma.com; c masnha@aero.tgp.ernet.n Abstract-Ths paper presents a meod for arcraft parameter estmaton usng neura senstvty anayss. The resuts are found to be superor to oer ANN based meods. I. INTRODUCTION The stabty and contro dervatves of an arcraft are of engneerng sgnfcance for mang fght smuator database, desgnng contro aw, epandng fght enveope and meetng arworness requrements. There are two broad approaches to estmate stabty and contro dervatves from fght data: drect approach and ndrect approach. Wdey apped parameter estmaton technques e output error meod, fter error meod and equaton error meod beong to e frst category. In ndrect approach, a non-near fter s used to estmate e parameters whch are defned as artfca state varabes []. Furermore, some meods can wor n rea tme (on-ne parameter estmaton and some dea w postfght data (off-ne parameter estmaton. However, ese conventona meods need nowedge about nta vaues of e parameters (often supped by computatona fud dynamcs (CFD smuatons and wnd tunne tests. Poor ntazaton may resut n dvergence of e agorm. Addtonay, conventona meods requre a pror nowedge of e aerodynamc mode (mode dentfcaton probem. Ths poses a severe mtaton to e appcabty of ese meods for compe aerodynamc stuatons e sta hysteress, arge-amptude tme-dependent maneuvers and hgh ange of attac fght where comng up w a reastc non-near aerodynamc mode s not so obvous [2]. The abty of artfca neura networ (ANN to act as unversa functon appromator offers sgnfcant advantages for arcraft parameter estmaton probem. It can capture hghy non-near compe phenomena n goba sense wout a pror nowedge about e dynamc mode. Aso nta estmates of e parameters are not necessary. These two quates render ANN a natura choce for estmatng stabty and contro dervatves from fght data. An eary appcaton of ANN for arcraft parameter estmaton can be found n [3] where moton and contro varabes were mapped to aerodynamc forces and moments usng bac propagaton agorm. Varous aspects of feed forward neura networ for dentfcaton of aerodynamc coeffcents were dscussed n deta n [4], [5] and [6]. Rasnghan, Ghosh and Kara [7] proposed two technques, caed deta meod and zero meod, for computng e stabty and contro TABLE I PARTICULARS OF THE MULTI-LAYER FEED FORWARD NEURAL NETWORK No. Attrbute Best Choce No. of Hdden Layers No. of Neurons (Hdden Hdden2 Output Actvaton Functon (Hdden Hdden2 Output Scang Range (Input & Output Inta Random Weghts dervatves usng feed forward neura networ w bac propagaton earnng. In e frst technque, e dervatves were nterpreted as varatons n e aerodynamc co-effcents due to a sma (deta varaton n one of e moton or contro varabes n such a fashon at ony at varabe undergoes ncrementa change whe e rest reman at er nomna vaues (hence e name deta meod. In e second technque, e dervatves were nterpreted as e rato of varaton n e vaue of e aerodynamc co-effcents to e ncrementa varaton n one of e moton or contro varabes whe e rest of em reman dentcay zero (hence e name zero meod. Bo e technques empoy centra dfference appromaton for computng e dervatves. In s paper, a nove neuro-computng approach s presented to address e probem of arcraft parameter estmaton by means of senstvty anayss of e aerodynamc forces and moments w respect to moton varabes and contro nputs. The net secton outnes e meodoogy. Smuaton resuts are presented net, foowed by concuson. II. METHODOLOGY The ratonae behnd appyng senstvty anayss for estmatng stabty and contro dervatves es n nterpretng em as e senstvty of e aerodynamc forces and moments on e moton and contro varabes. The moton and contro varabes are taen as e nput to e ANN and e force and moment co-effcents are taen as e output of e networ. A hgh (ow vaue of a parameter (dervatve mpes at output varabe of e ANN s hghy (weay senstve to e correspondng nput varabe. A mut-ayer feed forward neura networ (MFNN w two hdden ayers was used n s study for nput-output mappng. The partcuars of e networ are summarzed n Tabe I. The senstvty anayss of e ANN s dscussed net tanh tanh Identty -.9 to to +. IEEE ~ 229
2 Internatona Conference on Integent and Advanced Systems 27 Inputs Hdden ayer Hdden ayer 2 Output ayer Fg.. Neura networ archtecture The Neura networ mode used conssts of two hdden ayers and one output ayer as shown n fgure. The nput to e networ s desgnated as ~, whch s a vector of sze I + ncudng e bas. The nputs are fed to e frst hdden ayer; output of e frst hdden ayer s fed to e second hdden ayer; and output of e second hdden ayer s fed to e output ayer whch conssts of near neurons.e., ust summaton functons. The hdden ayers neurons consst of tangent hyperboc functon as a neurona nonnearty. The outputs of e neurons for varous ayers are desgnated as ~ y, ~ z, o~, as shown n e fgure. There are J number of neurons n e frst hdden ayer, K number of neurons n e second hdden ayer, and L number of neurons n e output ayer. Therefore, e foowng equatons can be wrtten for e outputs of e frst hdden ayer, second hdden ayer and e output ayer of neurons. For e frst hdden ayer, where, I y f + u f s e nput; u s e weght connectng nput to e neuron n e frst hdden ayer; f (. s e tangent hyperboc functon and s defned as I net + u ( (2 λ e f ( ; λ beng e steepness factor. λ + e Smary, e output of e second hdden ayer can be wrtten as J z f y v + f ( net (3 where a e notatons and subscrpts foow e same connotaton as epaned above. The output ayer output can be wrtten as K o + z w net (4 where e subscrpt vares from to L. Meod of bac propagaton w scaed conugate earnng s used to tran e neura networ. Once e networ s traned en t represents e functon whch s contaned n e nput-output data,.e. networ has dentfed e functon embedded n e data e e parta dervatves of e outputs w respect to e nputs represent e senstvty and can be derved as foows. o f ( net (5 snce ere s no nonnearty n e output ayer of neurons, erefore f, where 23 ~
3 Internatona Conference on Integent and Advanced Systems 27 and z y f f z K y J w v ( net ( u ( The nput data are normazed before use and erefore, equaton for e senstvty can be wrtten as o o o n o n n n where e subscrpt n refers to e normazed vaue. The foowng equaton s utzed to normaze e nput and e output o ma o on o mn + dff (2 o ma o mn smary n mn + dff ma ma mn Here, o n, n are e normazed output and nput respectvey. o ma and o mn stands for e mnmum and e mamum vaue of e output. Smar connotaton stands for ma and mn aso. dff stands for e range of normazaton used. In s case because e normazaton was done n e range [.9,.9] hence e vaue of dff s (.8. Therefore, n o o n dff dff o ma ma o mn mn An nvestgaton was done to reach e optma combnaton for e number of neurons n each hdden ayer. Tangent hyperboc functons were used as actvaton functons for (6 (7 (8 (9 ( ( (3 (4 (5 each of e neurons. A scaed conugate gradent agorm was used for fast supervsed earnng [8]. The fght data used to tran e networ was for ateradrectona dynamcs of Advanced Technoogy Testng Arcraft System (ATTAS from DLR, German Aerospace Center, Germany. The data was of tota 85 seconds duraton w e sampng tme of.4 second. There were ree dstnct maneuvers performed by e pot: short perod mode (for frst 25 seconds, usng mut-step eevator nput, ban-toban maneuver (for net 3 seconds, usng aeron nput and dutch ro (for ast 3 seconds, usng rudder doubet nput. Once traned, e MFNN was subected to testng data set, whch was taen to be e entre eng (228 sampes of tranng data set. Smuatons were performed for varyng number of teratons to assess e accuracy of mappng. The measured (fght derved force and moment co-effcents were en compared w ANN predctons. After predctng e aerodynamc co-effcents, e networ was used to estmate e stabty and contro dervatves usng senstvty anayss [9]. Une deta meod and zero meod, neura senstvty anayss computes e dervatves wout appyng fnte dfference appromaton. Snce parta dervatves are pont functons, e proposed meod yeds maematcay more accurate estmates. The superorty of e dervatves, estmated by e meod proposed here, s evdent from e resuts presented n e net secton. III. RESULTS AND OBSERVATIONS Numerous smuatons reveaed some mportant aspects of e MFNN. It was observed at reducng e number of neurons n e hdden ayer owers e order of e predcton curve resutng poor estmaton of e aerodynamc coeffcents. It was aso noted at usng a tangent hyperboc functon n e output ayer degrades e predcton. Hence ony weghted sum was performed at e output ayer. To mae a drect comparson, e same fght data was used to tran and test deta meod (whch gves better estmates an zero meod. Tabe II ceary shows at, smaer mean square error (MSE and cost functon vaue s acheved n ess number of teratons by scaed conugate gradent agorm used n s study. Beyond 5 teratons e MFNN shows no sgnfcant mprovement n predctons. Number of Iteratons Deta Meod TABLE II MSE AND COST FUNCTION MSE Senstvty Anayss Deta Meod Cost Functon Senstvty Anayss Startng 5.27e-3 6.7e-5 3.e e-3 Vaue e e-6 5.e-3.9e e-5 3.4e-7 4.7e-3.2e-9 5, 4.5e-5-3.8e-3 - In Fg. 2, 3 and 4, e measured (bue and ANN predcted (red aerodynamc co-effcents (sde force co-effcent, rong moment co-effcent and yawng moment co-effcent respectvey are potted w tme. These fgures show much superor tracng of e aerodynamc co-effcents compared ~ 23
4 Internatona Conference on Integent and Advanced Systems 27 to resuts obtaned by empoyng output error meod, fter error meod, equaton error meod and deta meod on e same fght data. Sde Force Co-effcent (C Y RMS Vaue of e Parameters (MSAS Sde force (C Y Dervatves C Yδr.2 C Yp C Yδa Iteratons C Yr C Yβ Fg.5. RMS vaues of sde force dervatves over e entre tranng set Rong Moment Co-effcent(C Tme (sec Fg. 2. Measured (bue and ANN predcted (red sde force co-effcent RMS Vaue of e Parameters (MSAS Rong moment (C Dervatves C p C δa C r Iteratons C δr C β Tme (sec Fg. 3. Measured (bue and ANN predcted (red rong moment co-effcent.2 Fg.6. RMS vaues of rong moment dervatves over e entre tranng set.35 Yawng Moment (C n Dervatves Yawng Moment Co-effcent (C n RMS Vaue of e Parameters (MSAS C nr C np C nδa C nδr C nβ Tme (sec Fg. 4. Measured (bue and ANN predcted (red yawng moment co-effcent Iteratons Fg.7. RMS vaues of yawng moment dervatves over e entre tranng set 232 ~
5 Internatona Conference on Integent and Advanced Systems C Ya C C nr Yp CYr C na C r C np C Yr C C Y C r C p Ca C n C nr Fg.8. Reatve standard devatons of e aerodynamc dervatves Durng tranng cyce, weghts of a networ get adusted. Once traned, e weghts of an ANN reman fed but e actvaton vaues of e neurons st change across e tranng set. Correspondng to each tranng vector one senstvty matr w resut. In s partcuar case, 228 sampes w resut 228 senstvty matrces. To get an average measure of e parameters over e entre tranng set, RMS vaues of e stabty and contro dervatves were obtaned. These vaues are often [9] termed as mean square average senstvtes (MSAS. Fg. 5, 6 and 7 depcts e convergence of e RMS vaues of e parameters (MSAS over e teratons. One woud epect e sde force to be more senstve to ange of sdesp and yaw rate and ess senstve to ro rate and aeron nput. Its senstvty on rudder nput shoud be moderate. These are ndeed e case as shown n Fg. 4. Agan rong moment shoud be hghy senstve on ro rate and aeron nput and weay senstve on ange of sdesp and rudder nput. Fg. 5 ustfes ese arguments. The sgnfcant dependence of rong moment co-effcent on yaw rate s a resut of ro-yaw coupng. Smary, e trends of e parameters shown n Fg. 6 corroborate we w ose epected from e nowedge of fght dynamcs. Thus e RMS vaues of e parameters not ony unfy e gradents of a e tranng vectors of e ANN, but aso provde an nsght to e mechancs of fght. The hstogram dstrbuton (not shown n s etended abstract of e parameters show near norma dstrbuton and hence er armetc means were taen as e average vaues of e parameters. superor estmaton of aerodynamc forces and moments compared to a estng meods for arcraft parameter estmaton usng neura networ. REFERENCES [] K. W. Iff, Parameter estmaton for fght vehce, J. Gudance, Contro and Dynamcs, vo. 2, no. 5, pp , 989. [2] R. V. Jategaonar, Fght Vehce System Identfcaton: A Tme Doman Meodoogy, st ed., vo. 26, Progress n Astronautcs and Aeronautcs, AIAA, Reston, VA, 26. [3] R. A. Hess, On e use of bac propagaton w feed forward neura networs for aerodynamc estmaton probem, AIAA Paper , August 993. [4] H. M. Youseff, Estmaton of aerodynamc coeffcents usng neura networs, AIAA Paper , August 993. [5] Basappa and R. V. Jategaonar, Aspects of feed forward neura networ modeng and ts appcaton to atera-drectona fght data, DLR-IB - 95/3, Braunschweg, Germany, September 995. [6] D. J. Lnse and R. F. Stenge, Identfcaton of aerodynamc coeffcents usng computatona neura networs, J. Gudance, Contro and Dynamcs, vo. 6, no. 6, pp.8-25, 993. [7] S. C. Rasnghan, A. K. Ghosh and P. K. Kara, Two new technques for parameter estmaton usng neura networs, The Aeronautca Journa, vo. 2, no., pp , 998. [8] M. F. Møer, A scaed conugate gradent agorm for fast supervsed earnng, Neura Networs, vo. 6, pp , 993. [9] J. M. Jurada, A. Manows and I. Coete, Senstvty anayss for mnmzaton of nput data dmenson for feedforward neura networ, IEEE Internatona Symposum on Crcuts and Systems, vo. 6, pp , 3 May-2 June, 994. IV. CONCLUSION A nove meod for estmatng arcraft stabty and contro dervatves s proposed n s paper based on neura senstvty anayss. The proposed meod shows much ~ 233
Neural network-based athletics performance prediction optimization model applied research
Avaabe onne www.jocpr.com Journa of Chemca and Pharmaceutca Research, 04, 6(6):8-5 Research Artce ISSN : 0975-784 CODEN(USA) : JCPRC5 Neura networ-based athetcs performance predcton optmzaton mode apped
More informationMARKOV CHAIN AND HIDDEN MARKOV MODEL
MARKOV CHAIN AND HIDDEN MARKOV MODEL JIAN ZHANG JIANZHAN@STAT.PURDUE.EDU Markov chan and hdden Markov mode are probaby the smpest modes whch can be used to mode sequenta data,.e. data sampes whch are not
More informationSupervised Learning. Neural Networks and Back-Propagation Learning. Credit Assignment Problem. Feedforward Network. Adaptive System.
Part 7: Neura Networ & earnng /2/05 Superved earnng Neura Networ and Bac-Propagaton earnng Produce dered output for tranng nput Generaze reaonaby & appropratey to other nput Good exampe: pattern recognton
More informationEEE 241: Linear Systems
EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they
More informationImage Classification Using EM And JE algorithms
Machne earnng project report Fa, 2 Xaojn Sh, jennfer@soe Image Cassfcaton Usng EM And JE agorthms Xaojn Sh Department of Computer Engneerng, Unversty of Caforna, Santa Cruz, CA, 9564 jennfer@soe.ucsc.edu
More informationResearch on Complex Networks Control Based on Fuzzy Integral Sliding Theory
Advanced Scence and Technoogy Letters Vo.83 (ISA 205), pp.60-65 http://dx.do.org/0.4257/ast.205.83.2 Research on Compex etworks Contro Based on Fuzzy Integra Sdng Theory Dongsheng Yang, Bngqng L, 2, He
More informationA finite difference method for heat equation in the unbounded domain
Internatona Conerence on Advanced ectronc Scence and Technoogy (AST 6) A nte derence method or heat equaton n the unbounded doman a Quan Zheng and Xn Zhao Coege o Scence North Chna nversty o Technoogy
More informationSampling-based Approach for Design Optimization in the Presence of Interval Variables
0 th Word Congress on Structura and Mutdscpnary Optmzaton May 9-4, 03, Orando, orda, USA Sampng-based Approach for Desgn Optmzaton n the Presence of nterva Varabes Davd Yoo and kn Lee Unversty of Connectcut,
More informationAssociative Memories
Assocatve Memores We consder now modes for unsupervsed earnng probems, caed auto-assocaton probems. Assocaton s the task of mappng patterns to patterns. In an assocatve memory the stmuus of an ncompete
More informationRESEARCH ARTICLE. Solving Polynomial Systems Using a Fast Adaptive Back Propagation-type Neural Network Algorithm
Juy 8, 6 8:57 Internatona Journa of Computer Mathematcs poynomas Internatona Journa of Computer Mathematcs Vo., No., Month, 9 RESEARCH ARTICLE Sovng Poynoma Systems Usng a Fast Adaptve Back Propagaton-type
More informationSupplementary Material: Learning Structured Weight Uncertainty in Bayesian Neural Networks
Shengyang Sun, Changyou Chen, Lawrence Carn Suppementary Matera: Learnng Structured Weght Uncertanty n Bayesan Neura Networks Shengyang Sun Changyou Chen Lawrence Carn Tsnghua Unversty Duke Unversty Duke
More informationThe Application of BP Neural Network principal component analysis in the Forecasting the Road Traffic Accident
ICTCT Extra Workshop, Bejng Proceedngs The Appcaton of BP Neura Network prncpa component anayss n Forecastng Road Traffc Accdent He Mng, GuoXucheng &LuGuangmng Transportaton Coege of Souast Unversty 07
More informationCOXREG. Estimation (1)
COXREG Cox (972) frst suggested the modes n whch factors reated to fetme have a mutpcatve effect on the hazard functon. These modes are caed proportona hazards (PH) modes. Under the proportona hazards
More informationA DIMENSION-REDUCTION METHOD FOR STOCHASTIC ANALYSIS SECOND-MOMENT ANALYSIS
A DIMESIO-REDUCTIO METHOD FOR STOCHASTIC AALYSIS SECOD-MOMET AALYSIS S. Rahman Department of Mechanca Engneerng and Center for Computer-Aded Desgn The Unversty of Iowa Iowa Cty, IA 52245 June 2003 OUTLIE
More informationCHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE
CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng
More informationNeural networks. Nuno Vasconcelos ECE Department, UCSD
Neural networs Nuno Vasconcelos ECE Department, UCSD Classfcaton a classfcaton problem has two types of varables e.g. X - vector of observatons (features) n the world Y - state (class) of the world x X
More informationREAL-TIME IMPACT FORCE IDENTIFICATION OF CFRP LAMINATED PLATES USING SOUND WAVES
8 TH INTERNATIONAL CONERENCE ON COMPOSITE MATERIALS REAL-TIME IMPACT ORCE IDENTIICATION O CRP LAMINATED PLATES USING SOUND WAVES S. Atobe *, H. Kobayash, N. Hu 3 and H. ukunaga Department of Aerospace
More information[WAVES] 1. Waves and wave forces. Definition of waves
1. Waves and forces Defnton of s In the smuatons on ong-crested s are consdered. The drecton of these s (μ) s defned as sketched beow n the goba co-ordnate sstem: North West East South The eevaton can
More information3. Stress-strain relationships of a composite layer
OM PO I O U P U N I V I Y O F W N ompostes ourse 8-9 Unversty of wente ng. &ech... tress-stran reatonshps of a composte ayer - Laurent Warnet & emo Aerman.. tress-stran reatonshps of a composte ayer Introducton
More informationExample: Suppose we want to build a classifier that recognizes WebPages of graduate students.
Exampe: Suppose we want to bud a cassfer that recognzes WebPages of graduate students. How can we fnd tranng data? We can browse the web and coect a sampe of WebPages of graduate students of varous unverstes.
More informationMultispectral Remote Sensing Image Classification Algorithm Based on Rough Set Theory
Proceedngs of the 2009 IEEE Internatona Conference on Systems Man and Cybernetcs San Antono TX USA - October 2009 Mutspectra Remote Sensng Image Cassfcaton Agorthm Based on Rough Set Theory Yng Wang Xaoyun
More informationNested case-control and case-cohort studies
Outne: Nested case-contro and case-cohort studes Ørnuf Borgan Department of Mathematcs Unversty of Oso NORBIS course Unversty of Oso 4-8 December 217 1 Radaton and breast cancer data Nested case contro
More informationBezier curves. Michael S. Floater. August 25, These notes provide an introduction to Bezier curves. i=0
Bezer curves Mchael S. Floater August 25, 211 These notes provde an ntroducton to Bezer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of the
More informationMACHINE APPLIED MACHINE LEARNING LEARNING. Gaussian Mixture Regression
11 MACHINE APPLIED MACHINE LEARNING LEARNING MACHINE LEARNING Gaussan Mture Regresson 22 MACHINE APPLIED MACHINE LEARNING LEARNING Bref summary of last week s lecture 33 MACHINE APPLIED MACHINE LEARNING
More informationGeneralized Linear Methods
Generalzed Lnear Methods 1 Introducton In the Ensemble Methods the general dea s that usng a combnaton of several weak learner one could make a better learner. More formally, assume that we have a set
More information1 GSW Iterative Techniques for y = Ax
1 for y = A I m gong to cheat here. here are a lot of teratve technques that can be used to solve the general case of a set of smultaneous equatons (wrtten n the matr form as y = A), but ths chapter sn
More informationThe Chaotic Robot Prediction by Neuro Fuzzy Algorithm (2) = θ (3) = ω. Asin. A v. Mana Tarjoman, Shaghayegh Zarei
The Chaotc Robot Predcton by Neuro Fuzzy Algorthm Mana Tarjoman, Shaghayegh Zare Abstract In ths paper an applcaton of the adaptve neurofuzzy nference system has been ntroduced to predct the behavor of
More informationApplication of support vector machine in health monitoring of plate structures
Appcaton of support vector machne n heath montorng of pate structures *Satsh Satpa 1), Yogesh Khandare ), Sauvk Banerjee 3) and Anrban Guha 4) 1), ), 4) Department of Mechanca Engneerng, Indan Insttute
More informationA Hybrid Learning Algorithm for Locally Recurrent Neural Networks
Contemporary Engneerng Scences, Vo. 11, 2018, no. 1, 1-13 HIKARI Ltd, www.m-hkar.com https://do.org/10.12988/ces.2018.711194 A Hybrd Learnng Agorthm for Locay Recurrent Neura Networks Dmtrs Varsams and
More informationShort-Term Load Forecasting for Electric Power Systems Using the PSO-SVR and FCM Clustering Techniques
Energes 20, 4, 73-84; do:0.3390/en40073 Artce OPEN ACCESS energes ISSN 996-073 www.mdp.com/journa/energes Short-Term Load Forecastng for Eectrc Power Systems Usng the PSO-SVR and FCM Custerng Technques
More information4DVAR, according to the name, is a four-dimensional variational method.
4D-Varatonal Data Assmlaton (4D-Var) 4DVAR, accordng to the name, s a four-dmensonal varatonal method. 4D-Var s actually a drect generalzaton of 3D-Var to handle observatons that are dstrbuted n tme. The
More information1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands
Content. Inference on Regresson Parameters a. Fndng Mean, s.d and covarance amongst estmates.. Confdence Intervals and Workng Hotellng Bands 3. Cochran s Theorem 4. General Lnear Testng 5. Measures of
More informationNONLINEAR SYSTEM IDENTIFICATION BASE ON FW-LSSVM
Journa of heoretca and Apped Informaton echnoogy th February 3. Vo. 48 No. 5-3 JAI & LLS. A rghts reserved. ISSN: 99-8645 www.jatt.org E-ISSN: 87-395 NONLINEAR SYSEM IDENIFICAION BASE ON FW-LSSVM, XIANFANG
More informationWeek 5: Neural Networks
Week 5: Neural Networks Instructor: Sergey Levne Neural Networks Summary In the prevous lecture, we saw how we can construct neural networks by extendng logstc regresson. Neural networks consst of multple
More informationADVANCED MACHINE LEARNING ADVANCED MACHINE LEARNING
1 ADVANCED ACHINE LEARNING ADVANCED ACHINE LEARNING Non-lnear regresson technques 2 ADVANCED ACHINE LEARNING Regresson: Prncple N ap N-dm. nput x to a contnuous output y. Learn a functon of the type: N
More informationNumerical Investigation of Power Tunability in Two-Section QD Superluminescent Diodes
Numerca Investgaton of Power Tunabty n Two-Secton QD Superumnescent Dodes Matta Rossett Paoo Bardea Ivo Montrosset POLITECNICO DI TORINO DELEN Summary 1. A smpfed mode for QD Super Lumnescent Dodes (SLD)
More informationBézier curves. Michael S. Floater. September 10, These notes provide an introduction to Bézier curves. i=0
Bézer curves Mchael S. Floater September 1, 215 These notes provde an ntroducton to Bézer curves. 1 Bernsten polynomals Recall that a real polynomal of a real varable x R, wth degree n, s a functon of
More informationThe Geometry of Logit and Probit
The Geometry of Logt and Probt Ths short note s meant as a supplement to Chapters and 3 of Spatal Models of Parlamentary Votng and the notaton and reference to fgures n the text below s to those two chapters.
More informationOptimum Selection Combining for M-QAM on Fading Channels
Optmum Seecton Combnng for M-QAM on Fadng Channes M. Surendra Raju, Ramesh Annavajjaa and A. Chockangam Insca Semconductors Inda Pvt. Ltd, Bangaore-56000, Inda Department of ECE, Unversty of Caforna, San
More informationThis column is a continuation of our previous column
Comparson of Goodness of Ft Statstcs for Lnear Regresson, Part II The authors contnue ther dscusson of the correlaton coeffcent n developng a calbraton for quanttatve analyss. Jerome Workman Jr. and Howard
More informationMultilayer Perceptron (MLP)
Multlayer Perceptron (MLP) Seungjn Cho Department of Computer Scence and Engneerng Pohang Unversty of Scence and Technology 77 Cheongam-ro, Nam-gu, Pohang 37673, Korea seungjn@postech.ac.kr 1 / 20 Outlne
More informationAdaptive LRBP Using Learning Automata for Neural Networks
Adaptve LRBP Usng Learnng Automata for eura etworks *B. MASHOUFI, *MOHAMMAD B. MEHAJ (#, *SAYED A. MOTAMEDI and **MOHAMMAD R. MEYBODI *Eectrca Engneerng Department **Computer Engneerng Department Amrkabr
More informationCyclic Codes BCH Codes
Cycc Codes BCH Codes Gaos Feds GF m A Gaos fed of m eements can be obtaned usng the symbos 0,, á, and the eements beng 0,, á, á, á 3 m,... so that fed F* s cosed under mutpcaton wth m eements. The operator
More informationQuantum Runge-Lenz Vector and the Hydrogen Atom, the hidden SO(4) symmetry
Quantum Runge-Lenz ector and the Hydrogen Atom, the hdden SO(4) symmetry Pasca Szrftgser and Edgardo S. Cheb-Terrab () Laboratore PhLAM, UMR CNRS 85, Unversté Le, F-59655, France () Mapesoft Let's consder
More informationNeuro-Adaptive Design - I:
Lecture 36 Neuro-Adaptve Desgn - I: A Robustfyng ool for Dynamc Inverson Desgn Dr. Radhakant Padh Asst. Professor Dept. of Aerospace Engneerng Indan Insttute of Scence - Bangalore Motvaton Perfect system
More informationNumerical integration in more dimensions part 2. Remo Minero
Numerca ntegraton n more dmensons part Remo Mnero Outne The roe of a mappng functon n mutdmensona ntegraton Gauss approach n more dmensons and quadrature rues Crtca anass of acceptabt of a gven quadrature
More informationFor now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.
Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson
More informationDISTRIBUTED PROCESSING OVER ADAPTIVE NETWORKS. Cassio G. Lopes and Ali H. Sayed
DISTRIBUTED PROCESSIG OVER ADAPTIVE ETWORKS Casso G Lopes and A H Sayed Department of Eectrca Engneerng Unversty of Caforna Los Angees, CA, 995 Ema: {casso, sayed@eeucaedu ABSTRACT Dstrbuted adaptve agorthms
More informationKernel Methods and SVMs Extension
Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general
More informationNUMERICAL DIFFERENTIATION
NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the
More informationLINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity
LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have
More informationLogistic Regression Maximum Likelihood Estimation
Harvard-MIT Dvson of Health Scences and Technology HST.951J: Medcal Decson Support, Fall 2005 Instructors: Professor Lucla Ohno-Machado and Professor Staal Vnterbo 6.873/HST.951 Medcal Decson Support Fall
More informationMultilayer Perceptrons and Backpropagation. Perceptrons. Recap: Perceptrons. Informatics 1 CG: Lecture 6. Mirella Lapata
Multlayer Perceptrons and Informatcs CG: Lecture 6 Mrella Lapata School of Informatcs Unversty of Ednburgh mlap@nf.ed.ac.uk Readng: Kevn Gurney s Introducton to Neural Networks, Chapters 5 6.5 January,
More informationWAVELET-BASED IMAGE COMPRESSION USING SUPPORT VECTOR MACHINE LEARNING AND ENCODING TECHNIQUES
WAVELE-BASED IMAGE COMPRESSION USING SUPPOR VECOR MACHINE LEARNING AND ENCODING ECHNIQUES Rakb Ahmed Gppsand Schoo of Computng and Informaton echnoogy Monash Unversty, Gppsand Campus Austraa. Rakb.Ahmed@nfotech.monash.edu.au
More informationApplication of Particle Swarm Optimization to Economic Dispatch Problem: Advantages and Disadvantages
Appcaton of Partce Swarm Optmzaton to Economc Dspatch Probem: Advantages and Dsadvantages Kwang Y. Lee, Feow, IEEE, and Jong-Bae Par, Member, IEEE Abstract--Ths paper summarzes the state-of-art partce
More informationPredicting Model of Traffic Volume Based on Grey-Markov
Vo. No. Modern Apped Scence Predctng Mode of Traffc Voume Based on Grey-Marov Ynpeng Zhang Zhengzhou Muncpa Engneerng Desgn & Research Insttute Zhengzhou 5005 Chna Abstract Grey-marov forecastng mode of
More informationMultigradient for Neural Networks for Equalizers 1
Multgradent for Neural Netorks for Equalzers 1 Chulhee ee, Jnook Go and Heeyoung Km Department of Electrcal and Electronc Engneerng Yonse Unversty 134 Shnchon-Dong, Seodaemun-Ku, Seoul 1-749, Korea ABSTRACT
More information9 Adaptive Soft K-Nearest-Neighbour Classifiers with Large Margin
9 Adaptve Soft -Nearest-Neghbour Cassfers wth Large argn Abstract- A nove cassfer s ntroduced to overcome the mtatons of the -NN cassfcaton systems. It estmates the posteror cass probabtes usng a oca Parzen
More informationLINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity
LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 31 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 6. Rdge regresson The OLSE s the best lnear unbased
More informationEbrahim Sharifi Tashnizi Tafresh University Dep. of Industrial and Mechanical Engineering Tafresh, Iran
Ebrahm Sharf ashnz et a. Ebrahm Sharf ashnz Dr.sharf@taut.ac.r afresh Unversty Dep. of Industra and Mechanca Engneerng afresh, Iran Azadeh Akhavan aher A.AkhavanaherBoroen@student.tudeft.n U Deft Unversty
More informationBoundary Value Problems. Lecture Objectives. Ch. 27
Boundar Vaue Probes Ch. 7 Lecture Obectves o understand the dfference between an nta vaue and boundar vaue ODE o be abe to understand when and how to app the shootng ethod and FD ethod. o understand what
More informationPsychology 282 Lecture #24 Outline Regression Diagnostics: Outliers
Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.
More informationOn the Power Function of the Likelihood Ratio Test for MANOVA
Journa of Mutvarate Anayss 8, 416 41 (00) do:10.1006/jmva.001.036 On the Power Functon of the Lkehood Rato Test for MANOVA Dua Kumar Bhaumk Unversty of South Aabama and Unversty of Inos at Chcago and Sanat
More informationChapter 13: Multiple Regression
Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to
More informationA MODIFIED METHOD FOR SOLVING SYSTEM OF NONLINEAR EQUATIONS
Journal of Mathematcs and Statstcs 9 (1): 4-8, 1 ISSN 1549-644 1 Scence Publcatons do:1.844/jmssp.1.4.8 Publshed Onlne 9 (1) 1 (http://www.thescpub.com/jmss.toc) A MODIFIED METHOD FOR SOLVING SYSTEM OF
More information1 Convex Optimization
Convex Optmzaton We wll consder convex optmzaton problems. Namely, mnmzaton problems where the objectve s convex (we assume no constrants for now). Such problems often arse n machne learnng. For example,
More informationDesign and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm
Desgn and Optmzaton of Fuzzy Controller for Inverse Pendulum System Usng Genetc Algorthm H. Mehraban A. Ashoor Unversty of Tehran Unversty of Tehran h.mehraban@ece.ut.ac.r a.ashoor@ece.ut.ac.r Abstract:
More informationSupporting Information
Supportng Informaton The neural network f n Eq. 1 s gven by: f x l = ReLU W atom x l + b atom, 2 where ReLU s the element-wse rectfed lnear unt, 21.e., ReLUx = max0, x, W atom R d d s the weght matrx to
More informationDownlink Power Allocation for CoMP-NOMA in Multi-Cell Networks
Downn Power Aocaton for CoMP-NOMA n Mut-Ce Networs Md Shpon A, Eram Hossan, Arafat A-Dwe, and Dong In Km arxv:80.0498v [eess.sp] 6 Dec 207 Abstract Ths wor consders the probem of dynamc power aocaton n
More informationSPATIAL KINEMATICS OF GEARS IN ABSOLUTE COORDINATES
SATIAL KINEMATICS OF GEARS IN ABSOLUTE COORDINATES Dmtry Vasenko and Roand Kasper Insttute of Mobe Systems (IMS) Otto-von-Guercke-Unversty Magdeburg D-39016, Magdeburg, Germany E-ma: Dmtr.Vasenko@ovgu.de
More informationSimulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests
Smulated of the Cramér-von Mses Goodness-of-Ft Tests Steele, M., Chaselng, J. and 3 Hurst, C. School of Mathematcal and Physcal Scences, James Cook Unversty, Australan School of Envronmental Studes, Grffth
More informationMODEL TUNING WITH THE USE OF HEURISTIC-FREE GMDH (GROUP METHOD OF DATA HANDLING) NETWORKS
MODEL TUNING WITH THE USE OF HEURISTIC-FREE (GROUP METHOD OF DATA HANDLING) NETWORKS M.C. Schrver (), E.J.H. Kerchoffs (), P.J. Water (), K.D. Saman () () Rswaterstaat Drecte Zeeand () Deft Unversty of
More informationREAL-TIME DETERMINATION OF INDOOR CONTAMINANT SOURCE LOCATION AND STRENGTH, PART II: WITH TWO SENSORS. Beijing , China,
REAL-TIME DETERMIATIO OF IDOOR COTAMIAT SOURCE LOCATIO AD STREGTH, PART II: WITH TWO SESORS Hao Ca,, Xantng L, Wedng Long 3 Department of Buldng Scence, School of Archtecture, Tsnghua Unversty Bejng 84,
More informationChapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems
Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons
More informationLogistic Regression. CAP 5610: Machine Learning Instructor: Guo-Jun QI
Logstc Regresson CAP 561: achne Learnng Instructor: Guo-Jun QI Bayes Classfer: A Generatve model odel the posteror dstrbuton P(Y X) Estmate class-condtonal dstrbuton P(X Y) for each Y Estmate pror dstrbuton
More informationTypical Neuron Error Back-Propagation
x Mutayer Notaton 1 2 2 2 y Notaton ayer of neuron abeed 1,, N neuron n ayer = vector of output from neuron n ayer nput ayer = x (the nput pattern) output ayer = y (the actua output) = weght between ayer
More informationChapter 14 Simple Linear Regression
Chapter 4 Smple Lnear Regresson Chapter 4 - Smple Lnear Regresson Manageral decsons often are based on the relatonshp between two or more varables. Regresson analss can be used to develop an equaton showng
More informationRegularized Discriminant Analysis for Face Recognition
1 Regularzed Dscrmnant Analyss for Face Recognton Itz Pma, Mayer Aladem Department of Electrcal and Computer Engneerng, Ben-Guron Unversty of the Negev P.O.Box 653, Beer-Sheva, 845, Israel. Abstract Ths
More informationGrover s Algorithm + Quantum Zeno Effect + Vaidman
Grover s Algorthm + Quantum Zeno Effect + Vadman CS 294-2 Bomb 10/12/04 Fall 2004 Lecture 11 Grover s algorthm Recall that Grover s algorthm for searchng over a space of sze wors as follows: consder the
More informationPattern Classification
Pattern Classfcaton All materals n these sldes ere taken from Pattern Classfcaton (nd ed) by R. O. Duda, P. E. Hart and D. G. Stork, John Wley & Sons, 000 th the permsson of the authors and the publsher
More informationSystem Identification for Quad-rotor Parameters Using Neural Network
EVERGREE Jont Journal of ovel Carbon Resource Scences & Green Asa Strategy, Volume 3, Issue, pp. 6-, March 6 System Identfcaton for Quad-rotor Parameters Usng eural etwor Tare. Def,*, Shgeo Yoshda Graduate
More informationNetworked Cooperative Distributed Model Predictive Control Based on State Observer
Apped Mathematcs, 6, 7, 48-64 ubshed Onne June 6 n ScRes. http://www.scrp.org/journa/am http://dx.do.org/.436/am.6.73 Networed Cooperatve Dstrbuted Mode redctve Contro Based on State Observer Ba Su, Yanan
More informationCSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography
CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve
More informationNon-Linear Back-propagation: Doing. Back-Propagation without Derivatives of. the Activation Function. John Hertz.
Non-Lnear Bac-propagaton: Dong Bac-Propagaton wthout Dervatves of the Actvaton Functon. John Hertz Nordta, Begdamsvej 7, 200 Copenhagen, Denmar Ema: hertz@nordta.d Anders Krogh Eectroncs Insttute, Technca
More informationReliability Sensitivity Algorithm Based on Stratified Importance Sampling Method for Multiple Failure Modes Systems
Chnese Journa o Aeronautcs 3(010) 660-669 Chnese Journa o Aeronautcs www.esever.com/ocate/ca Reabty Senstvty Agorthm Based on Strated Importance Sampng Method or Mutpe aure Modes Systems Zhang eng a, u
More informationLinear Approximation with Regularization and Moving Least Squares
Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...
More informationCIS526: Machine Learning Lecture 3 (Sept 16, 2003) Linear Regression. Preparation help: Xiaoying Huang. x 1 θ 1 output... θ M x M
CIS56: achne Learnng Lecture 3 (Sept 6, 003) Preparaton help: Xaoyng Huang Lnear Regresson Lnear regresson can be represented by a functonal form: f(; θ) = θ 0 0 +θ + + θ = θ = 0 ote: 0 s a dummy attrbute
More informationAn identification algorithm of model kinetic parameters of the interfacial layer growth in fiber composites
IOP Conference Seres: Materals Scence and Engneerng PAPER OPE ACCESS An dentfcaton algorthm of model knetc parameters of the nterfacal layer growth n fber compostes o cte ths artcle: V Zubov et al 216
More informationGrid Generation around a Cylinder by Complex Potential Functions
Research Journal of Appled Scences, Engneerng and Technolog 4(): 53-535, 0 ISSN: 040-7467 Mawell Scentfc Organzaton, 0 Submtted: December 0, 0 Accepted: Januar, 0 Publshed: June 0, 0 Grd Generaton around
More informationDynamic Analysis Of An Off-Road Vehicle Frame
Proceedngs of the 8th WSEAS Int. Conf. on NON-LINEAR ANALYSIS, NON-LINEAR SYSTEMS AND CHAOS Dnamc Anass Of An Off-Road Vehce Frame ŞTEFAN TABACU, NICOLAE DORU STĂNESCU, ION TABACU Automotve Department,
More informationCorrespondence. Performance Evaluation for MAP State Estimate Fusion I. INTRODUCTION
Correspondence Performance Evauaton for MAP State Estmate Fuson Ths paper presents a quanttatve performance evauaton method for the maxmum a posteror (MAP) state estmate fuson agorthm. Under dea condtons
More informationIV. Performance Optimization
IV. Performance Optmzaton A. Steepest descent algorthm defnton how to set up bounds on learnng rate mnmzaton n a lne (varyng learnng rate) momentum learnng examples B. Newton s method defnton Gauss-Newton
More informationDr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur
Analyss of Varance and Desgn of Exerments-I MODULE III LECTURE - 2 EXPERIMENTAL DESIGN MODELS Dr. Shalabh Deartment of Mathematcs and Statstcs Indan Insttute of Technology Kanur 2 We consder the models
More informationβ0 + β1xi. You are interested in estimating the unknown parameters β
Revsed: v3 Ordnar Least Squares (OLS): Smple Lnear Regresson (SLR) Analtcs The SLR Setup Sample Statstcs Ordnar Least Squares (OLS): FOCs and SOCs Back to OLS and Sample Statstcs Predctons (and Resduals)
More informationCorrelation and Regression. Correlation 9.1. Correlation. Chapter 9
Chapter 9 Correlaton and Regresson 9. Correlaton Correlaton A correlaton s a relatonshp between two varables. The data can be represented b the ordered pars (, ) where s the ndependent (or eplanator) varable,
More informationFault Diagnosis of Autonomous Underwater Vehicles
Research Journal of Appled Scences, Engneerng and Technology 5(6): 407-4076, 03 SSN: 040-7459; e-ssn: 040-7467 Maxwell Scentfc Organzaton, 03 Submtted: March 3, 0 Accepted: January, 03 Publshed: Aprl 30,
More informationWhy feed-forward networks are in a bad shape
Why feed-forward networks are n a bad shape Patrck van der Smagt, Gerd Hrznger Insttute of Robotcs and System Dynamcs German Aerospace Center (DLR Oberpfaffenhofen) 82230 Wesslng, GERMANY emal smagt@dlr.de
More informationOptimization of JK Flip Flop Layout with Minimal Average Power of Consumption based on ACOR, Fuzzy-ACOR, GA, and Fuzzy-GA
Journa of mathematcs and computer Scence 4 (05) - 5 Optmzaton of JK Fp Fop Layout wth Mnma Average Power of Consumpton based on ACOR, Fuzzy-ACOR, GA, and Fuzzy-GA Farshd Kevanan *,, A Yekta *,, Nasser
More informationOne-sided finite-difference approximations suitable for use with Richardson extrapolation
Journal of Computatonal Physcs 219 (2006) 13 20 Short note One-sded fnte-dfference approxmatons sutable for use wth Rchardson extrapolaton Kumar Rahul, S.N. Bhattacharyya * Department of Mechancal Engneerng,
More informationChapter 9: Statistical Inference and the Relationship between Two Variables
Chapter 9: Statstcal Inference and the Relatonshp between Two Varables Key Words The Regresson Model The Sample Regresson Equaton The Pearson Correlaton Coeffcent Learnng Outcomes After studyng ths chapter,
More information