This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

N V R Kishore Ms. S Mondal Dr. R Kar Dr. D Mandal Prof.S P Ghoshal

Objective:

Minimize the fitness function by using Improved Particle Swarm Optimization algorithm for designing of a BS FIR filter.

Specification of the filter requirements. Calculation of suitable filter coefficients.

Methodology: Representation of the filter by a suitable structure. (realization) Analysis of the effects of finite word length on filter performance. Implementation of filter in software and/or hardware.

Problem Formulation The frequency response of the FIR digital filter can be calculated as. 2T k wk ! N . H e where.

! § h .

The frequency is sampled in [0.n e jwk n !0 N jwk n . The expression for error function E . ] with N points.

[ ! G .

[ « H d e j[ H i e j[ » ½ .

.

j[ Where H d e is the frequency response of the designed filter and for .

band stop filter it is given as. H d e j[k ! 0 .

!1 otherwise and H i . for [cl e [ e [ch .

The error to be minimized is defined as J1 ! max E .e j[ is the frequency response of the ideal filter.

[ H p max E .

[ H s [ e[ p .

[ u[ s .

.

y Sampling number : 128 yOrder of the Coefficient : 20 yNumber of iteration : 40 .85 for BS. 3.75. Benchmarks assume: y Cut off frequency : lower edge frequency 0.5 version on core (TM) 2 duo processor. y Sampling frequency : 1Hz.35 and upper edge frequency 0.25. 0. yMATLAB 7.Experimental Setup y Platform: Microsoft Windows platform .00GHz with 2 GB RAM. 0.

This is called recombination. This results in completely new trait in the organism. These rules are encoded in genes of the organism which are connected by long strings called chromosomes. Very occasionally a gene may be mutated. The fittest organism is more probable to survive. . Two organisms mate to share their genes. Resultant offspring may end up having half genes of one parent & half of the other.REAL CODED GENETIC ALGORITHM(RGA) Genetic Algorithm(GA) is mainly a probabilistic search technique based on the principles of natural selection & evolution. Every organism has a set of rules describing the characteristics of the organism.

The iteration stops when the maximum number of cycles is reached. . Genetic cycle updating. Crossover and mutation to generate off-springs. The grand minimum and its corresponding chromosome string or the desired solution are finally obtained.Real coded Genetic Algorithm (RGA) Steps of RGA as implemented for optimization of current excitations are: y Initialization of real chromosome strings of np population. Size of the set depends on the number of excitation elements in a particular filter design. Copying of the elite strings over the non-selected strings. Decoding of strings and evaluation of error of each string. Selection of elite strings in order of increasing values from the minimum value. each y y y y y y consisting of a set of current excitations.

. inheritance. and crossover". This heuristic is routinely used to generate useful solutions to optimization and search problems. which generate solutions to optimization problems using techniques inspired by natural evolution. Genetic algorithms belong to the larger class of evolutionary algorithms (EA). such as mutation. selection."The genetic algorithm (GA) is a search heuristic that mimics the process of natural evolution.

Bird flocking optimizes a certain objective function. PSO concept is similar to the behaviour of a swarm of birds. which can easily handle with non-differential objective functions. unlike traditional optimization methods. . It was developed in 1995 by Russell Eberhart & James Kennedy.PARTICLE SWARM OPTIMIZATION (PSO) PSO is a robust stochastic optimization technique based on the movement and intelligence of swarms with implicit parallelism. PSO is developed through simulation of bird flocking in multidimensional space.

Each agent tries to modify its position using the following Information: y 1. Mathematically. robust population based stochastic search technique. 2. Each agent knows its best value so far (pbest). y Each agent knows the best value so far in the group (gbest) among pbests. velocities of the particles are modified according to Vi .Standard Particle Swarm Optimization (PSO) PSO is a flexible. The distance between the current position and gbest. The distance between the current position and pbest.

k 1 ! w Vi k C1 rand1 pbestik Sik .

C 2 rand 2 gbest k Sik .

(1) And the searching point in the solution space may be modified by Si.

k 1 ! Sik Vi .

k 1 (2) .

The two random variables rand1. The modified PSO is termed as IPSO. which has a memory of the previous worst. the good experience component. rand2 are replaced by one single variable r the other being 1-r. The velocity expression of the PSO is split into two components viz.IMPROVED PARTICLE SWARM OPTIMIZATION (IPSO) The global search ability of traditional PSO is very much enhanced the help of following modifications. . which has a memory of its previously visited best position & the bad experience component. The addition of bad experience component to the particle gives an additional exploration capacity to the swarm.

one single random number r is chosen so that when r1 is large.Improved Particle Swarm Optimization (IPSO) For IPSO instead of taking independent rand1 and rand 2 . .

sign. Another random number r2 is introduced. r1 1 1 is small and vice versa.

Finally the modified velocity becomes Vi .r3 is introduced for considering reverse velocity.

k 1 ! r2 sign.

r3 Vi k .

r2 C1 r 1 pbest ik S k 1 i .

r2 C 2 .

r1 gbest k S k (1 r2 ) * c1 * r1 S k pworst ik 1 1 i i _ a _ .

a where sign.

r3 ! 1 !1 where r3 e 0.05 where r3 " 0.05 .

05 2.0 0.0 - .0 1.4 IPSO 25 200 2.01 1.01 1.01 Gaussian Mutation Roulette 1/3 wmax wmin PSO 25 350 2.Results and discussions The parameter values of RGA and PSO and IPSO are given in the following tables.05 0.05 2.05 0. Parameters Population size Iteration Cycle Crossover rate Crossover Mutation rate Mutation Selection Selection Probability C1 C2 RGA 120 600 1 Two Point Crossover 0.

095808566461517 -0.003567579531894 -0.118 0.028502857888104 -0.400369877077545 PSO 0.002300832301629 -0.082 0.0939 0.001893868108392 -0.096334206843426 0.005029715870387 0.075465671961181 -0.055176711318501 0.04054 0.055687299039737 -0.000020082854816 0.297783056747784 0.001915220607266 0.72 30.298190578247113 -0.001324681739036 0.02966 Transition width Execution Time Per 100 cycles 2.002697869505629 0.076189026154460 0.004760681054475 -0.0941 0.3379 0.0909 0.025165091322598 -0.000639149080848 0.100057194730152 0.001409980793664 -0.03262 0.65 23.81 29.001962570281435 0.Table : Optimized Coefficients of FIR Band stop Filter of Order 20 h(N) h(1)=h(21) h(2)=h(20) h(3)=h(19) h(4)=h(18) h(5)=h(17) h(6)=h(16) h(7)=h(15) h(8)=h(14) h(9)=h(13) h(10)=h(12) h(11) RGA 0.400369877077545 IPSO 0.125 BS filter Maximum stop band ripple (normalized) 0.053196793860741 -0.003939814376923 -0.128 0.0936 .0177 2.299380312728113 -0.56 Maximum pass band ripple (normalized) 0.8389 2.000752480372393 0.000994123920259 0.072788608489910 0.400369877077545 Table : Summery of IPSO results with other algorithms for Band stop filter of Order 20 Algorithm Maximum stop band attenuation (dB) PM RGA PSO IPSO 21.027902253292513 0.08273 0.

3 0.4 0.8 0.2 0.1 0. .7 0.5 0.6 F requ en cy 0.9 1 Figure: Magnitude (dB) Plot of the FIR Band stop Filter of Order 20.0 -10 M a g n itu d e (d B ) -20 -30 -40 -50 -60 -70 0 PM R GA PSO IP S O 0.

. Fig: Convergence Profile for IPSO in case of 20th Order Band stop FIR Filters.9 8 10 8 7 6 E rro r 5 4 3 2 1 0 100 200 300 400 Iteration C ycles 500 600 0 0 20 40 60 80 100 120 Iteration C ycles 140 160 180 200 E rro r 4 2 6 Fig: Convergence Profile for RGA in case of 20th Order BS FIR Filter.

Conclusions IPSO outperforms RGA and conventional PSO in accuracy of magnitude response. . IPSO converges to a much lower fitness in lesser number of iterations .

this can be used for designing IIR filter.Future Scope The proposed method is simple and the computational complexity is less so. . Try to involve Frequency Response Masking (FRM) technique with this optimizing methods in future work. Only the magnitude and error fitness concept is considered. the delay concept will be considered in future work.

pp. Audio Electro acoust. Audio Electro acoust. Wiley.R.¶ IEEE International Symposium on Circuits and Systems. J.¶ Signal Processing. IEEE International Symposium on Circuits and Systems.4. Antoniou A. Schussler. O.W. N.151±155. L. Parks..S Swamy.: µDesigning digital FIR filters using Tabu search algorithm. Ahmad SU. Gonos. 80. Kalinli A. Mastorakis. Lett.: µA genetic algorithm approach for fractional delay FIR filters¶. 1997. IEEE Trans. pp. 497-505 S. Approximate design relationships for High-pass FIR digital filters. IIR Model Identification Using Batch-Recursive Adaptive Simulated Annealing Algorithm. Parks. Chebyshev approximation for non recursive digital filters with linear phase. pp.H. I.H.Bibliography y y y y y y y y y y y L. McClellan. vol.F. Rabiner. 0278-6648.2236-2239 . 2000.W. Karaboga D. M. Shian-Tang Tzeng. New York.E.N. T. Electron. Herrmann. T.S. McClellan. Karaboga N. Burrus. Litwin. W. In Proceedings of 6th Annual Chinese Automation and Computer Science Conference. Horrocks DH. 1987. IEEE Trans. Circuits Theory. pp.W. Chen.: µDesign of arbitrary FIR log filters by genetic algorithm approach. A computer program for designing optimum FIR linear phase digital filters. ISCAS 2006. L. Rabiner. 2000.. C.2517-2520 Hung-Ching Lu.Fundamental Theory and Applications. Design of non-recursive digital filters with linear phase.. 6 (1970). 50 (2003) 634±639. Parks. FIR and IIR digital filters. ISCAS '97. AU-21 (1973) 506±526. 2000. IEEE Transaction on Circuits and Systems I . CT-19 (1972) 189±194. Design of Two Dimensional Recursive Filters Using Genetic Algorithms. 28±31. Digital Filter Design. J. IEEE Potentials. IEEE Trans.R. AU-21 (1973) 456±460. 329±330. T.

Panigrahi. Kennedy. H. pp. H. 56. 38.. Venayagamoorthy. S. Mandal. 18. 346.´ IEEE Trans.328±348 B. pp. and A. no. Expert Systems with Applications. . DEPSO and PSO-QI in digital filter design. Linear phase FIR filter design using particle swarm optimization and genetic algorithms. 3954-3961. H.y Karaboga N. S. Chan. (4). Bataineh. Digital Signal Processing. 2008. Ghoshal. ³Power quality disturbance classification using fuzzy Cmeans algorithm and adaptive particle swarm optimization. Ghoshal. H. and A. Conf. G. Expert Systems with Applications. 657±668. 38 (2011).K. 1995. Differential Evolution Particle Swarm Optimization for Digital Filter Design. K. D. Y. 2009. 30±39. Comparative Optimal Designs of Non-uniformly Excited Concentric Circular Antenna Array Using Evolutionary Optimization Techniques. Eberhart. y y y y y y y y y y Journal of the Franklin Institute. F. and K. pp. R. Ababneh. Conf.. Bhattacharjee. ³Improved hybrid particle swarm optimized wavelet neural network for modeling the development of fluid dispensing for electronic packaging. Bhattacharjee. 2008. P. 2010.I. Particle Swarm Optimization. Luitel. Eberhart. 619-624. Bhattacharjee. PP. K. Iu. Rabi Kumar Mahapatra. M. 55. Proc. Leung. 3. 2010. D.´ IEEE Trans. 2000. 1. Comparison between Genetic Algorithms and Particle Swarm Optimization. Mandal. K. Application of Evolutionary Optimization Techniques for Finding the Optimal set of Concentric Circular Antenna Array. IEEE int. Ghoshal. Siba Prasada Panigrahi. ³Swarm Intelligence Based Optimal Design of Concentric Circular Antenna Array. no. S. 2009. H. 3447±3460. Shi. Y.´ Journal of Electrical Engineering. C. 9. 7th Ann. and A. R. Sep. K. Dash. vol. Ind. Archana Sarangi. P. F. vol. Jan. 2008. Ling. On Neural Network. J. (Elsevier). San Diego. vol. pp. S. Biswal. P. ICETET¶09 (2009). Electron. Mandal. K. 2942-2950. 2008 IEEE Congress on Evolutionary Computation (CEC 2008). in Proc. J. and B. 10966±10973. IEEE Second International Conference on Emerging Trends in Engineering and Technology. pp. on Evolutionary Computation. Electron. vol. P. D. 212±220. B. no.: µA new design method based on artificial bee colony algorithm for digital IIR filters¶. Ind. 10.

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd