Comparison Of Different Techniques To Design Of Filter
International Journal of Computer Applications (0975 – 8887)Volume 97– No.1, July 2014Comparison of Different Techniques to Design of FilterShushank DograNarinder SharmaDepartment of Electronics and communicationAmritsar college of Engeering and Technology,Amritsar, IndiaDepartment of Electronics and CommunicationAmritsar college of Engeering and Technology,Amritsar, IndiaABSTRACTFor the design of filters complex calculations are required.Mathematically, by substituting the values of parametersaccording to any of the filter used in any of the methods fromwindow method, frequency sampling method or optimalmethod we can get the values of filter coefficients h(n).Thereare different window methods like Kaiser window,rectangularwindow,triangular window ect.Kaiser window method hasbeen chosen preferably because of the presence of ripplefactor (β).Here, I have design Band pass FIR and low passfilter using artificial neural network which gives optimumresult i.e. the difference between the actual and desired outputis minimum and also shows the comparision between lowpass and bandpass filter or comparison of different techniquesused to design the filter.1. INTRODUCTIONThe basic function of digital filter is to eliminate the noise andto extract the signal of interest from other signals. A digitalfilter filter is a basic device used in digital signal processing.There are several techniques available to design the digitalfilters. But generally while designing a digital filter, first ananalog filter is designed and then it is converted into thecorresponding digital filter. With the technological evolution,great advances have been made on design techniques forvarious digital filters. A filter is essentially a system ornetwork that selectively changes the wave shape amplitude –frequency and or phase – frequency characteristics of a signalin a desired manner . A digital filter is a mathematicalalgorithm implemented in hardware and/or software thatoperates on a digital input signal to produce a digital outputsignal for the purpose of achieving a filtering objective.(Wade et al., 1990) described filters are synthesized bycascading elements from a library of computationally simpleand some case very computationally efficient, primitive filter.A simplified block diagram of a real-time digital filter, withanalog input and output signals is given in Figure 1.1 Cut-off frequency: This is frequency which separatespass band and stop band.Types of analog filters:The different types of analog filters are as follows: Low pass filter (L.P.F): It passes the frequency from0 upto some designated frequency, called as cut-offfrequency. After cut-off frequency, it will not allowany signal to pass through it.High pass filter (H.P.F): It passes the frequencyabove some designated frequency called as cut-offfrequency. If input signal frequency is less than the cutoff frequency, then this signal is not allowed to passthrough it.Band Pass Filter (B.P.F): It allows the frequenciesbetween two designated cut-off frequencies.Band Stop Filter (B.S.F): It attenuates all frequenciesbetween two designated cut-off frequencies, while itpasses all other frequencies.All Pass Filter: It passes all the frequencies. But byusing this filter the phase of input signal can bemodified.Types of digital filters: FIR and IIR filters[Ifeachor]Digital filters are broadly divided into two classes, namelyinfinite Impulse Response (IIR) and Finite Impulse Response(FIR) filters. Either type of filter, in its basic form, can berepresented by its impulse response sequence, h(k) (k 0, 1, 2 .), as inFig.2.A conceptual representation of a digital filterThe equation for IIR isFig 1: Block diagram of digital filterA.Analog FiltersThis is necessary because generally digital filters are designedusing analog filters. Some parameters related to analog filters: The equation for FIR isPass band: It passes certain range of frequencies. Inthe pass band, attenuation is 0.Stop band: It suppresses certain range of frequencies.In the stop band, attenuation is infinity.25
International Journal of Computer Applications (0975 – 8887)Volume 97– No.1, July 2014Merits of FIR filters over IIR filters FIR filters have an exactly linear phase responses where thephase responses of IIR filters are non linear. FIR are realized nonrecursively, that is by direct evaluationare always stable. The stability of IIR filters cannot always beguaranteed. The coefficient quantization error is much less severe in FIRthan in IIR.2. STEPS OF FIR FILTER DESIGNThe design of a digital filter involves five steps as shown infigure:zeros, making it appear as though the waveform suddenlyturns on and off:W(n) 1The M point rectangular window, which corresponds to directtruncation of the Fourier series, has the weighted functionWr (n) 1,for-(M-1) /2 n (M-1)/20,otherwiseTriangular windowTriangular windows are given by:where L can be N, N 1, or N-1. The latter is also knownas Bartlett window. All three definitions converge at large N.The triangular window is the 2nd order B-spline window andcan be seen as the convolution of two half-sized rectangularwindows, giving it twice the width of the regular windows.Bartlett WindowBy tapering the rectangular window sequence linearly fromthe middle to the ends, we then obtain the M point triangularwindow given byWT (n) 2n /(M-1)Fig.3.summary of design stages of filterThe criteria is a linear phase response in frequency domaincalled phase response (Jin et al., 2006) as shown in Fig.Finally, because there is a tradeoff between filter ndimplementation feasibility, complexity is a performancecriteria. Ideal filter characteristics are practically unrealizable.A.METHODS TO DESIGN FIR FILTERWe have many methods to design FIR filter that are: Fourier series methodfor 0 n (M-1)/22-2n / (M-1)for (M-1)/2 n (M-1)0,otherwiseThis window is also called Barlett window.Raised Cosine WindowsAnother class of windows is raised cosine windows that,compared with the triangular window, are smoother at theends, but closer to one at the middle. The smoother taper atthe ends should reduce the side-lobe levels, while the broadermiddle section reduces distortion of the desired pulse responsebound n 0 (Garimella, 2008). To reduce the side-lobe levelfurther, we can consider an even more gradual taper at theends of the window sequence by using the raised cosinesequence. The various windows in this category are: Frequency Sampling method Window methodThe most of these design techniques suffer from some kindsof drawback, Some of them could not give optimal design inany sense, some is lacking of generality, some need longcomputing time, and so on (Bagachi and Mitra, 1996). Kaiserwindow method has been used because of the presence ofripple parameter beta.1. Hanning windowThe Hann window named after Julius von Hann and alsoknown as the Hanning (for being similar in name and form tothe Hamming window), von Hann and the raised cosinewindow is defined byWc (n) 3. TYPES OF WINDOW FUNCTIONSThere are many windows proposed that approximate thedesired characteristics. The basic window functions are listedbelow:2. Hamming WindowHamming window is given byRectangular WindowTherectangularwindow(sometimesknownasthe boxcar or Dirichlet window) is the simplest window,equivalent to replacing all but N values of a data sequence by26
International Journal of Computer Applications (0975 – 8887)Volume 97– No.1, July 2014follows connectionist architecture, and parallel distributedprocessing. Artificial Neural Networks are collections ofmathematical models that emulate some of the observedproperties of biological nervous system. The key element ofthe ANN is the novel structure of the information processingsystems.3. Blackman windowBlackman window is given byWb(n) Some other advanrtages of ANN are as under:The difference between these windows are shown below:TABLE.1 Comparison of WindowsWindowTypePeak Sidelobeamplitude(relative)ApproximateWidth ofmain lobe1.2.3.Adaptive learningSelf-OrganisationReal Time OperationA.Multilayer Perceptron Networks[3],[4]PeakApproximationError 20logδ(dB)Rectangular-134π/M π/M-53Blackmann-5712π/M-74Multilayer Perceptron Networks [3-4]form a class of feedforward neural networks. They are not a single layer networkbut consist of an input layer, arbitrary number of hiddenlayers and an output layer as shown in figure 1.5.Here input isfed to each of the input layer neurons. The outputs of the inputlayer feed into each of next layer neurons and so on, forminga layered structure having one input layer , one output layerand L-2 hidden layers in an L layer network.The followingdiagram illustrates a perceptron network with three Layers:4. KAISER WIDOWKaiser determined empirically that the value of β need toachieve a specified value of A is given by 0.1102 A 8.7 0.4 0.5842 A 21 0.07886 A 21 0.0 for A 50for 21 A 50for A 21Recall that the case β 0 is the rectangular window for whichA 21. Further more, to achieve prescribed values of A anddf, M must satisfy equations A 8 1 M 14.36df 0.922 / df 1 for A 21for A 21Finite Impulse response filters (Öner and paper., 1999) arepreferred for their stable and linear phase characteristics. Butdue to long impulse response of FIR filters there will be morehardware complexity.5. ARTIFICIAL NEURAL NETWORKAn Artificial Neural Network is an information processingparadigm inspired by the way the densely inter-connected,parallel structure of the mammalian brain process information.ANN have successfully applied to a number of problemsincluding the identification and control of dynamical systems,communications networks, coordination of robotics handeyemovements.It is also referred to as a neuromorphic system,Fig.2: Multilayer PerceptronB. Formulation of ProblemThe design of digital filter means basically finding the valuesof filter coefficients so that given filter specification areachieved the window based design method are exclusivelyused for calculating there coefficient. We have used Kaiserwindow for this purpose. The Kaiser window function goessomehow in overcoming the incorporating a ripple controlparameter, ANNs have been used for the design of Bandpassfilter with Passband ripple, stop band attenuation, passbandfrequency F1, passband frequency F2, sampling frequencyand for lowpass with Passband, transition width, passbandripple, stop band attenuation, sampling frequency as inputparameters.The network has been trained in such a manner sothat the error comes minimum, means there may be very lessdifference in the results comes from actual calculations thathas been come from matlab and the output comes from trainedartificial neural network.C.OBJECTIVE:The objective of the present work are divided into thefollowing sections.(1)To prepare the data sheet using different values of filterparameter achieve the filter coefficient.(2)Choosing ANN a Band-pass FIR filter has been designedsuch that its coefficient match with coefficientsobtainedwith window method.27
International Journal of Computer Applications (0975 – 8887)Volume 97– No.1, July 20146. COMPARISON OF RESULTS OFLOWPASS FILTER WITHBANDPASS FILTERFilter coefficient are calculated and works is carried out usingapproximately 30 such values of all the above parameters tocalculate the filter coefficients. The range of differentparameters of lowpass and bandpass has been taken whichare:TABLE.2.Comparison of lowpass with bandpass filterLOWPASSFILTERPARAMETERSBT: (10.1040– 9.3860)TW: (50 -100)PBR: (0.1- 0.2)N: (3–10)PBF:(150- 180)BANDPASSFILTERPARAMETERAp: (0.7-1.3) dbAs: (40-55) dbF1: (7000-11000) hzF2: (17000-21000) hzFs:(47000-52000) hzUsing this data set the Artificial Neural Network has beentrained and can be use to calculate filter coefficients for inputparameters in this range.Now, ANN is use to design the bandpass and lowpass FIR filter.There is very less difference in theann results and the calculated results.INPUT PARAMETER OF LOWPASS- Transitionwidth 50Hz,SamplingFrequency 1KHz,Passbandripple 0.1db,Filterlength 3,Passband 150Hz,stopbandattenuation 10.1040TABLE.3.Comparison of Kaiser window method )h(10)h(11)h(12)Kaiser WindowMethodArtificialNeuralNetwork7. CONCLUSIONThe present work has illustrated the need of Artificial NeuralNetwork for the design of Filters and also shows thecomparision of all other methods to design the filter.The following conclusions are drawn from this research work: Artificial Neural Network is better and easy method ofDesign of FIR Filter. Using Fourier series, Frequency sampling or Windowmethods the filter can be design but for each unknownparameter the filter coefficients have to calculated. Incomparison with ANN, the trained network can calculate thefilter Coefficient for unknown parameter in that specifiedrange.0.28850.28058. 000.0200[1] V. Aggarwal, J. O. Wesley and M. O Una., (2006) “Filterapproximation using Explicit Time and FrequencyRemain specification”. Proceeding, of the AnnualSymposium on Artificial Intelligence, 2006, Seattle,Washington, pp. 0.0000INPUT PARAMETER OF BANDPASS- Passband ripple 0.729,Stopband attenuation 41.502 ,PassbandfrequencyF1 9815,PassbandfrequencyF2 18234 ,Samplingfrequency 49661. Filter Length 17TABLE.4. Error values obtained from difference betweenKaiser window and ANN.[2] A. M. Cristian and H. Guinther, (2007) “ArchitectureOptimization of a Finite Impulse Response Filter usingtogglebased power estimation”, International conference,on intelligent and Advance systems, 2007, pp. 1270–1273.[3] A. Fizelow., P. Brites., A. Ochoa., H. Mertuns., E.Fernandez, and Garcia-Martinez R. (2007) “FindingOptimal Neural Network Architecture using GeneticAlgorithms,” Proceeding of the advances in computerand Engineering, 2007, pp. 15–24.[4] D. Bhattacharya and A. Antoniou, (1996) “Real TimeDesign of FIR Filters of Feedback Neural Network”,Vol. 3, 1996, pp. 1070–1078[5] D. A. Yasur and R. Teresa., (2007) “LUT-based PowerMacro modeling Technique for DSP architectures”,Proceeding of the IEEE, Centro de Electronica Industrial,Spain, August, 2007, pp. 1416–1419.28
International Journal of Computer Applications (0975 – 8887)Volume 97– No.1, July 2014[6] K.B. Englihart, B.S. Hudgins, M. Stevenson and P.A.Parker, (1994) “Myoelectric Signal Classification using aFinite Impulse Response Neural Network”. TechnicalReport, The university of New Brunswork, Canada, June,1994, pp. 803–820.[7] Z. Fan and P. Mars, (1997) “Access flow control schemefor ATM networks using neural-network-based trafficprediction”. Proceeding for IEEE, Vol. 144, Dec. 5,October 1997, pp. 708– 714.[8] G. Rama Marthy. (2008) “Finite Impulse Response FIRfilter Model of Synapses: Associated Neural Network”,Proceeding of the Fourth Annual IEEE InternationalConfidences on Natural Computation, 2008, pp. 3304–3309.[9] G. Zichao and U. E. Robert., (1992) “Using GeneticAlgorithms to select input for Neural Network”.Proceeding of the IEEE International conference, May,1992, pp. 87–95.[10] K.J. Hintz. and J.J.Spofford , (1990) “Evoling NeuralNetwork” Proceedings of the IEEE Transactions onCommunication and Intelligence, May 1990, pp. 333–338.IJCATM : www.ijcaonline.org[11] S. Haykins, (2003), “Neural Networks – Acomprehensive foundation”, Prentice – Hall of IndiaPrivate Limited, NewDelhi 2003.[12] I. Ioan., R. Corina and I. Arpad., (2004) “Theoptimization of feedforward”, proceedings of theinternational and information – ICTAMI 2004, Thessaloniki, Greece.[13] I F. Emmonual C. and J. Barriel W. (2001) “DigitalSignal Processing,” A Practical Approch,” PersonEducation (Singapore) Ltd., 2001, Second Edition.[14] J. Y. Dar. and C. F. Kun., (2007) “Least square Design ofFIR Filters based on a compacted Feedback NeuralNetwork”, Proceeding of the IEEE Transaction onCircuits and systems, vol. 54 issue 5, May, 2007, pp.427–431. “Magnetization, M”, not just “M”. If includingunits in the label, present them within parentheses. Donot label axes only with units. In the example, write“Magnetization (A/m)” or “Magnetization {A[m(1)]}”,not just “A/m”. Do not label axes with a ratio ofquantities and units. For example, write “Temperature(K)”, not “Temperature/K”.29
pass and bandpass filter or comparison of different techniques used to design the filter. 1. INTRODUCTION The basic function of digital filter is to eliminate the noise and to extract the signal of interest from other signals. A digital filter filter is a basic device used in digital signal processing.
work/products (Beading, Candles, Carving, Food Products, Soap, Weaving, etc.) ⃝I understand that if my work contains Indigenous visual representation that it is a reflection of the Indigenous culture of my native region. ⃝To the best of my knowledge, my work/products fall within Craft Council standards and expectations with respect to
Comparison of Singulation Techniques Electronic Packaging Society, Silicon Valley Chapter Sept. 28, 2017 ANNETTE TENG Sept 28, 2017 1 . Comparison Sept 28, 2017 Annetteteng@promex-ind.com 35 Plasma dicing TLS Sidewall Comparison of Saw Techniques Plasma Stealth Laser Ablation Laser
Undoubtedly, this research will enrich greatly the study on forecasting techniques for apparel sales and it is helpful to identify and select benchmark forecasting techniques for different data patterns. 2. Methodology for forecasting performance comparison This research will investigate the performances of different types of forecasting techniques
A COMPARISON OF SLIP MEASUREMENT TECHNIQUES. Raja Bharath Vangapattu, M.S. Department of Electrical Engineering Northern Illinois University, 2016 Dr. Donald S. Zinger, Director This thesis mainly deals with the slip frequency measurement of an induction motor using two slip measurement techniques and comparison of the techniques used.
Comparison table descriptions 8 Water bill comparison summary (table 3) 10 Wastewater bill comparison summary (table 4) 11 Combined bill comparison summary (table 5) 12 Water bill comparison – Phoenix Metro chart 13 Water bill comparison – Southwest Region chart 14
figure 8.29 sqt comparison map: superior bay (top of sediment, 0-0.5 ft) figure 8.30 sqt comparison map: 21st avenue bay figure 8.31 sqt comparison map: agp slip figure 8.32 sqt comparison map: azcon slip figure 8.33 sqt comparison map: boat landing figure 8.34 sqt comparison map: cargill slip figure
chart no. title page no. 1 age distribution 55 2 sex distribution 56 3 weight distribution 57 4 comparison of asa 58 5 comparison of mpc 59 6 comparison of trends of heart rate 61 7 comparison of trends of systolic blood pressure 64 8 comparison of trends of diastolic blood pressure 68 9 comparison of trends of mean arterial pressure
Water bill comparison summary (table 3) 10 Wastewater bill comparison summary (table 4) 11 Combined bill comparison summary (table 5) 12 Water bill comparison - Phoenix Metro chart 13 Water bill comparison - Southwest Region chart 14 Water bill comparison - 20 largest US cities chart 15
