Letter Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No.

2, July 2012

Application of Particle Swarm Optimization for the Design of IIR Digital Filters
Ranjit Singh1, and Sandeep K.Arya2
1

JMIT/ECE Dept., Yamuna Nagar, INDIA Email: chauhan443@yahoo.co.in 2 GJUS&T/ECE Dept., Hisar, INDIA algorithm (GA) [6][9], simulated annealing (SA), tabu search (TS) [5] etc. Analytical methods or Simple iterative usually lead to sub-optimal designs. Consequently, there is a need of optimization methods (heuristic type) that can be use to design digital filters that would satisfy prescribed specifications. Swarm intelligence has become a research interest to many research scientists from various areas in recent years. The swarm intelligence can be defined as any attempt for designing algorithms or distributed problemsolving devices inspired by the collective behaviour of insects and other animal societies. Particle Swarm Optimization, which was originally proposed by J.Kennedy and R.Eberhart [8], is a novel algorithm inspired by birds flocking in the sky or fish schooling. The drawback of preceding design methods is that the computation time is quite long. The Proposed algorithm has strong search capability and is superior to the SA and GA. The proposed algorithm is implemented in MATAB and results are found to be very encouraging. The outline of this paper is as follows: In Section II, a brief introduction to Particle Swarm Optimization (PSO) approach is discussed. Particle Swarm Optimization related to IIR digital filter design is proposed in section III. Two designed examples are used to verify the proposed method and simulation results of experiments are reported in Section IV. The conclusion and future scope is described in end of this paper. II. PARTICLE SWARM OPTIMIZATION Particle Swarm Optimization (PSO) is an evolutionary algorithm developed by Eberhart and Kennedy in 1995 [9]. It is a population based search algorithm and is inspired by the observation of natural habits of bird flocking and fish schooling. PSO is a flexible, robust population-based stochastic search/optimization technique with implicit parallelism, which can easily handle with non-differential objective functions, unlike traditional optimization methods. It is developed through simulation of bird flocking in multidimensional space. Bird flocking optimizes a certain objective function. Each particle knows its best value so far (pbest). This information corresponds to personal experiences of each particle. Moreover, each particle knows the best value so far in the group (gbest) among pbests. Namely, each particle tries to modify its position using the following information: The distance between the current position and pbest. The distance between the current position and gbest. 99

Abstract In this paper, Particle Swarm Optimization (PSO) in the context of designing infinite-impulse response (IIR) digital filters is presented. IIR filter is essentially a digital filter with Recursive responses. Since the error surface of digital IIR filters is generally nonlinear and multimodal, global optimization techniques are required in order to avoid local minima. This Paper presents heuristic way for the designing IIR filters. PSO is a powerful global optimization algorithm introduced in combinatorial optimization problems. The paper finds the optimum Coefficients of IIR digital filter through PSO. It is found that the calculated values are more optimal than fda and GA tool availble for the design of filter in MATLAB. Design of Lowpass and High pass IIR digital filter is proposed to provide estimate of transition band. The simulation results of the employed examples shows an improvement on transition band. The position of the PoleZero is also presented to describe the stability of designed filters. Index TermsDigital Signal Processing (DSP), Digital Filter, Infinite-Impulse Response (IIR), Optimization, Particle Swarm Optimization (PSO).

I. INTRODUCTION The use of Digital filters in implementing high performance circuits in digital signal environment is becoming increasingly popular. Digital filters are capable of performance specifications that would, at best, be extremely difficult, if not impossible, to achieve with an analog implementation. In addition, the characteristics of a digital filter can be easily changed under software control. Digital filters are classified either as Finite duration impulse response (FIR) filters or Infinite duration impulse response (IIR) filters, depending on the form of impulse response of the system. In the FIR system, the impulse response sequence is of finite duration, i.e., it has a finite number of non zero terms. Digital infinite-impulseresponse (IIR) filters can often provide a much better performance and less computational cost than their equivalent finite-impulse-response (FIR) filters [3] and have become the target of growing interest [1], [2], [4]. Digital IIR filters are widely used in the fields of automatic control, telecommunications, speech processing and etc. However, because the error surface of IIR filters is usually nonlinear and multimodal, conventional gradient-based design methods may easily get stuck [4], [5]. Therefore, some researchers have attempted to develop design methods based on modern heuristic optimization algorithms such as particle swarm optimization (PSO), ant colony optimization (ACO), genetic 2012 ACEEE DOI: 01.IJRTET.7.2.507

Letter Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 2, July 2012
L

Each individual is treated as a volume-less particle in the Ddimensional space, with the position and velocity of ith particle represented as Xi=(Xi1,Xi2,.,XiD) and Vi=(Vi1,Vi2,,ViD). The particles move according to the following equation: Vid = w*Vid + c1* r1*( Pid X id ) + c 2* r2* ( Pg X id ) (1) X id = X id + Vid (2) where c1 and c2 are acceleration coefficients, r1 and r2 are two random numbers. Vector Pi= (Pi1,Pi2,,PiD) is the best previous position (the position giving the best fitness value) of particle i called pbest, and vector Pg=(Pg1,Pg2, ,PgD) is the position of the best particle among all the particles in the population and called gbest. Parameter w is the inertia weight introduced to accelerate the convergence speed of the PSO.The process for implementing PSO is as follow: Initialize a population (array) of particles with random position and velocities on d dimensions in the problem space. For each particle, evaluate the desired optimization fitness function in d variables. Compare particles fitness evaluation with particles pbest. If current value is better then pbest, then set pbest location equal to the current location in d dimension space. Compare fitness evaluation with the populations overall previous best. If current value is better then gbest, then reset gbest to the current particles array index and value. Change the value and position of the particle according to above equations. Loop to step 2, until a criterion is met. The pseudo code for PSO is shown in fig.1.

A z H z B z

i0

a i z i
M

i 1

bi z i

(4)

These parameters a0, a1, a2, ., aL, b1, b2, .., bM appearing in eq. 3 and eq. 4 are called the filter coefficients which gives the characteristics of the filter. The flow chart for the Design of Digital filter is shown in fig. 2. An important task for the designer is to find values of ai and bi such that the frequency response of the filter approximates a desired characteristic while preserving the stability of the designed filter. Hence, the design of this filter can be considered as an optimization problem of cost function J(w) stated as min J(w) where w=[ a0, a1, a2, ., aL, b1, b2, .., bM ] is filter coefficient vector. The aim is to minimize the cost function J(w) by adjusting w.

Figure 2. Flow chart of Digital filter design

The cost function is usually expressed by:

2 N J (w ) 1 d (k ) y (k ) N k 1

(5)

Figure 1. Pseudo code of PSO

III. PSO FOR DESIGNING IIR DIGITAL FILTERS In general, Digital filters are classified as Recursive and Non-Recursive filters. The response of Recursive or IIR filters is dependent on one or more of its past output. IIR filters have lot of advantages over Non-Recursive filters. Consider the IIR filter with the input-output relationship governed by:

where d(k) and y(k) are the desired and actual responses of the filter, respectively and N is the number of samples. Particle Swarm Optimization based Infinite Impulse response (PSOIIR) digital filter is implemented by MATLAB. The filter coefficients were achieved using following parameters: with c1 and c2 each equal to 2, and the population size of 100 was assumed. w often is decreased linearly from 2 to 0.1 in 300 iterations. IV. RESULTS AND DISCUSSION Experiments are carried for well known IIR digital filters (see Appendix A), which has been used by many authors. The examples were performed on with Pentium IV, 2.80 GHz CPU and 1GB of RAM. The magnitude and phase response of first example is shown in fig. 3 and fig. 4 in which a zoomed in curve for transition band is included. 100

y k

b y k i a x k i
i i i 1 i0

(3)

where x(k) and y(k) are the filters input and output, respectively, M (e L) is the filter order. The transfer function of this IIR filter can be written as: 2012 ACEEE DOI: 01.IJRTET.7.2.507

Letter Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 2, July 2012

Figure 3. Magnitude Response of Low pass filter

Figure 4. Phase Response of Low pass filter

Figure 5. Pole-Zero Position of Low pass filter using: (a) PSO; (b) fda.

In fig. 5, we have summarized Pole-Zero behavior of Low pass filter which shows that the designed filter is stable using proposed method. The mean-square-error (MSE) and resulting values of the low-pass filter in term of coefficients are shown in Table 1 and Table 2. It is observed from table I
TABLE I. RESULTS
OF

MSE OF LOW-PASS FILTER

TABLE II. C OEFFICIENTS OF LOW-PASS FILTER DESIGNED WITH 3RD ORDER

Fig. 6 Magnitude Response of High Pass filter

that the proposed method gives optimal values of coefficients with minimum mean-square-error (MSE). Table 2 provides a comparison of coefficients with the filter design tool and Genetic Algorithm (GA). The fig. 6 and fig. 7 illustrates the magnitude and phase response of second example.

Fig. 7 Phase Response of High Pass filter

Fig. 8 summarized Pole-Zero position of High pass filter using fda and PSO which gives us stable high pass filter.

2012 ACEEE DOI: 01.IJRTET.7.2.507

101

Letter Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 2, July 2012

Fig. 8. Pole-Zero Position of High pass filter using: (a) PSO; (b) fda TABLE III. RESULTS OF MSE
OF

HIGH-PASS FILTER

response specifications and multiband. Further, the other Evolutionary algorithm or mematic algorithm can be discussed for the design of IIR filters. APPENDIX A

Table III evaluates the values of MSE for filters coefficients. Table 4 gives the coefficients of high-pass filter under same specifications.
TABLE IV. C OEFFICIENTS OF HIGH-PASS FILTER DESIGNED WITH 5TH ORDER

Example 1: In the first example, design a low-pass filter with following specifications: Pass/Stop band ripples 1dB/ 15dB, and band edges 200Hz / 400Hz and a sampling frequency of 1000Hz. Example 2: This example is taken for design of a highpass filter with following specifications: Pass/Stop band ripples 1dB/ 75dB, and band edges 600Hz / 200Hz and a sampling frequency of 1500Hz. REFERENCES
[1] David E. Goldberg, Genetic Algorithm in search, optimization and machine Learning, Pearson Education, Low price Edition, Delhi, 2005. [2] Andreas Antoniou, Digital Filters analysis, design and application, Tata Mcgraw-hill edition, New Delhi, 2005. [3] A.P. Engelbrecht, Fundamentals of Computational Swarm Intelligence, John Wiley & Sons Ltd., New Delhi, 2005. [4] Joelle Skaf and P. Boyd Stephen, Filter Design with Low Complexity Coefficients, IEEE Transactions on Signal processing, vol. 56, no. 7, pp. 3162-3170, July 2008. [5] Richard J. Vaccaro and Brian F. Harrison, Optimal MatrixFilter Design, IEEE Transactions on Signal processing, vol. 44, no. 3, pp. 705-710, March 1996. [6] Xi Zhang and Hiroshi Iwakura, Design of IIR Digital Filters based on Eigen value Problem, IEEE Transactions on Signal processing, vol. 44, no. 6, pp. 1325-1319, June 1996. [7] Fabrizio Argenti and Enrico Del Re, Design of IIR Eigen filters in the Frequency domain, IEEE Transactions on Signal processing, vol. 46, no. 6, pp. 1694-1700, June 1998. [8] Xin Yao, Yong Liu, and Guangming Lin, Evolutionary Programming Made Faster, IEEE Transaction on Evolutionary Computation, vol. 3, no. 2, pp. 83-102, July 1999. [9] R.S.Chauhan and Sandeep K. Arya, IIR digital filter design through Genetic Algorithm, Elixir Advance Engg. Informatics Journal, vol. 30, no. 1, pp. 1796-1799, 2011.

CONCLUSION It is concluded that Particle Swarm Optimization is a global optimization technique for IIR digital filters, and the benefits of PSO for designing digital filter have been studied. The simulation results show that PSO has better, or at least equivalent, global search ability and convergence speed than others. The examples demonstrate the versatility of the proposed approach. The performance of proposed method has been compared with fda tool and GA. Better transition band is calculated from the proposed method. The position of Pole-Zero of the designed filters using optimal coefficients gives us stable filters. Thus it is believed that the proposed algorithm is capable of quick and high performance. The proposed method can be extended to arbitrary magnitude 2012 ACEEE DOI: 01.IJRTET.7.2.507 102