IARJSET ISSN (Online) 2393-8021 ISSN (Print) 2394-1588 International Advanced Research Journal in Science, Engineering and Technology Vol. 3, Issue 5, May 2016 Copyright to IARJSET DOI 10.17148/IARJSET.2016.3520 91 Design of FIR Filter Using Particle Swarm Optimization A.Praneeth 1 , Prashant K.Shah 2 M.Tech Student, Department of Electronics Engineering, NIT Surat, India 1 Associate Professor, Department of Electronics Engineering, NIT Surat, India 2 Abstract: This paper presents an optimal design of digital low pass finite impulse response (FIR) filter using Particle Swarm Optimization (PSO). The design target of FIR filter is to approximate the ideal filters on the request of a given designing specifications. The traditional based optimization techniques are not efficient for digital filter design. The filter specification to be realized PSO algorithm generatesthe best coefficients and try to meet the ideal frequency response. Particle swarm optimization (PSO) proposes a new equation for the velocity vector and updating the particle vectors and hence the solution quality is improved. The PSO technique enhances its search capability that leads to a higher probability of obtaining the optimal solution. In this paper for the given problem the realization of the FIR filter has been performed. The simulation results have been performed by using the particle swarm optimization (PSO) method. Keywords: FIR Filter, PSO, GA, Parks and McClellan (PM) Algorithm. I. INTRODUCTION A filter is defined as a system that passes the band of frequencies according to the requirement. The aim of filtering is to improve the quality of the signal by removing unwanted component of signal such as noise. Different types of filters that are available are low pass filter, notch filters, high pass filter, band pass and band rejects filters etc. Filters are also classified as analog and digital. Analog filters are designed with various electronic components for example resistor, capacitor etc. and continuous time signal is used as an input. Digital filters are used in a various number of applications such as speech processing and image processing etc. The digital filter have number of advantages are: Digital filters can be used at low frequency, the frequency response can be changed as per the requirement if it was implemented by using a programmable processor and several input signals can be filtered by one digital filter without the need to replicate hardware. The digital filters disadvantages are: speed limitation and long design and development times. Depending on the form of the unit impulse response sequence digital filters may be divided into two categories FIR and IIR filters. FIR stands for finite impulse response filter that is its impulse response is of a finite duration and IIR stands for infinite impulse response filter which is defined as impulse response is of infinite duration. FIR filters have various number of advantages over IIR filters: FIR filters can have an exact linear phase, finite impulse response filters are stable, the design methods are generally linear, they can be realized effectively in hardware, No feedback is required which reduces the circuit complexity and hence fir filters are always stable. There are different methods for the design of fir filter for example window design techniques, frequency sampling method, optimal design methods and evolutionary algorithm techniques. For the filter design the aim is to find the filter coefficients by optimizing the error function. Window design method is the widely used method for the filter design. Some of the window functions used are Hamming window, Kaiser Window, Hanning window and Bartlet window. The Window function converts the infinite length response into the finite length response. Linear phase FIR filters are required when the time domain specifications are given [1]. The frequently used method for the design of linear phase weighted Chebyshev FIR digital is based on the Remez exchange algorithm by Parks and McClellan [2]. Further improvements to their results are reported in [3]. The limitation of this procedure is that the relative values of the amplitude error in the frequency bands are specified by means of weighting function and not by the deviations themselves. A different evolutionary algorithm such as genetic algorithm (GA), Differential evolution and artificial bee colony optimization [5-6] etc. has been used for the design of digital filters. Although the GA have a good performance for finding the promising regions of the search space they are inefficient in determining the local minimum in terms of the convergence speed and solution quality [7]. This present paper the design of FIR filter using the evolutionary algorithm Particle Swarm Optimization (PSO). The PSO advantages lie in its simplicity to implement as well as its convergence can be controlled by few parameters. Some of the works done in order to explore the flexibility of FIR filter design provided by PSO [8-9]. PSO algorithm generates the best coefficients that try to meet ideal frequency characteristics. The PSO is simple technique to implement and its convergence may be controlled via few parameters. This paper is explained as follows: Section II includes the problem statement. Section III which includes Particle Swarm Optimization Algorithm. Section IV includes the results and analysis. Section V includes the conclusion and the reference.