DESIGN AND PERFORMANCE ANALYSIS OF BAND PASS IIR FILTER FOR SONAR APPLICATION

Transcription

1 International Journal of Emerging Tehnologies and Engineering (IJETE) ISSN: ICRTIET-21 Conferene Proeeding, 3 th -31 st August DESIGN AND PERFORMANCE ANALYSIS OF BAND PASS IIR FILTER FOR SONAR APPLICATION 1 Vanshikha Singh, 2 Rajesh Mehra 1 M.E.Sholar, 2 Assoiate Professor 1,2 Departmant of Eletronis & Communiation Engineering NITTTR, Chandigarh, India ABSTRACT In reent years due to the magnifient development of filter designs, designers took attention in this researh area. Filters are used to separate or ombine different frequenies. In this paper IIR filter has been designed and simulated using different tehniques., shev1, shev2 and filters are the basi prototype filter from whih we designed IIR filter. The performane of all the developed filters has been analyzed and ompared by using different filter order in MATLAB. It an be observed from the simulated result that the performane of filter is better at lower order and shev2 and filter shows better performane with inrease in filter order in SONAR frequeny range. Key Words: Band pass,, shev1, shev2,, IIR, SONAR. 1. INTRODUCTION ADSP is the proess of mathematial analysis of a signal to modify and improve its harateristis. Digital signal proessing (DSP) refers to various tehniques for improving the auray and reliability of digital ommuniations. The theory behind DSP is quite omplex. Basially, DSP works by larifying the levels or states of a digital signal. ADSP iruit is able to differentiate between human-made signals, whih are orderly, and noise, whih is inherently haoti [1]. The working of digital signal proessor is as follows: Fig. 1 Digital signal proessing The digital signal proessing is shown in Fig. 1 where we have to onvert the analog signal into digital signal by using A/D onverter and then the funtion of digital proessor starts. After transmission of digital signal suessfully at the reeiver terminal we need D/A onverter to onvert the digital signal into its analogous form. The basi task of DSP is to transform, filter and ompress the analog signal. After doing the task we are able to store information signal more effiiently. The advantages to using DSP tehniques are Reproduibility, Programmability, Stability and High reliability et. The appliations of DSP tehniques are Audio signal proessing and ompression, Digital image proessing video ompression, Speeh proessing and reognition, RADAR, SONAR Finanial signal proessing, Biomediine et. The basi task of DSP is filtering. Filtering is the one of the most powerful and important tool of DSP. Digital filters are apable of performing that speifiations whih are extremely diffiult, to ahieve with an analog implementation. In addition, the harateristis of a digital filter an be easily hanged under software ontrol. Many digital systems use signal filtering to remove unwanted noise, to provide spetral shaping, or to perform signal detetion or analysis. Digital filter appliations inlude signal preonditioning, band seletion, and low pass filtering. These funtions are provided by two types of filters i.e. Finite impulse response (FIR) filters and Infinite impulse response (IIR) filters. The basi harateristis of Finite Impulse Response (FIR) filters are Linear phase, high filter order (more omplex iruits) and Stability. The basi harateristis of Infinite Impulse Response (IIR) are Non-linear phase, Low filter order (less omplex iruits) and Resulting digital filter has the potential to beome unstable. IIR filters have muh better frequeny response than FIR filters of the same order. Unlike FIR filters, their phase is not linear whih an reate a problem to the systems

2 International Journal of Emerging Tehnologies and Engineering (IJETE) ISSN: ICRTIET-21 Conferene Proeeding, 3 th -31 st August whih need phase linearity. On the other hand, if the linear phase harateristi is not required, the use of IIR filters is an exellent solution [2]. 2. IIR FILTER IIR filters are digital filters with infinite impulse response. Unlike FIR filters, they have the feedbak (a reursive part of a filter) and are known as reursive digital filters. For this reason IIR filters have muh better frequeny response than FIR filters of the same order. When the linear phase harateristi is not important, the use of IIR filters is an exellent solution [2]. performane. After filter speifiation we have to alulate the oeffiient a k and b k and realize the struture to analyze finite word length and then hardware and software implementation[2].there are four types of basi prototype filter from whih we design IIR filter[]. The filter has a maximally flat response, i.e, No pass bands ripples and roll-off of minus 2db per pole [6]. Low pass filters are all-pole filter haraterized by the magnitude-squared frequeny response (3) Where N is the order of the filter, Ω is its -3dB frequeny, Ω p is the pass band edge frequeny, and is the band edge value of. Fig. 2 Blok Diagram of IIR Filter Fig. 2 shows the representation of IIR Filter in whih it onsist a feedbak element whih provides a good output response. There are many tehniques for designed of digital filter having a infinite duration impulse response. Thus the design of an IIR filter involves design of a digital filter in the analog domain and transforming the design into the digital domain. The system funtion desribing an analog filter may be written as: H a (s) = (1) Where a k and b k are the oeffiient of filter. The system funtion H a (s) an also be written by its impulse response as: For the IIR Filter design it is neessary to find to find suitable values for the oeffiients a k and b k so that the desired frequeny response is obtained. To design digital IIR filter first we have filter speifiation. The filter speifiation gives the funtion of the filter (low pass, high pass, band pass, band rejet filter) and the desired (2) There are two types of shev filters. Chebvshev1 filters are all-pole filters that exhibit equiripple behavior in the pass band and a monotoni harateristi in the stop band. The shev2 filters ontain both poles and zeros and exhibits an equiripple behavior in the stop band and a monotoni harateristi in the pass band. The zeros of type 2 filters lie on the imaginary axis in the s plane. The magnitude squared of the frequeny response harateristi of a shev1 filter is given as Where Ɛ is the parameter related to the pass band ripple and T N (x) is the Nth order shev polynomial defined as () T N (x) = os(nos -1 x), () osh (Nos -1 x), The magnitude squared of the frequeny response harateristi of a shev2 filter is given as (6) Where T N (x) is the Nth order hebyshev polynomial and Ω s is the stop band frequeny. or Cauer filters exhibit equiripple behavior in both the pass band and the stop band. This lass of filters ontains both poles and zeros and is haraterized by the magnitude squared frequeny response.

3 Group delay (in samples) Group delay (in samples) International Journal of Emerging Tehnologies and Engineering (IJETE) ISSN: ICRTIET-21 Conferene Proeeding, 3 th -31 st August (7) Where U N is the Jaobian funtion of order N and Ɛ is the parameter related to the pass band ripple [3]. 3. IIR DESIGN SIMULATION The Magnitude response and the group delay of the band pass filter using, shev1, shev2 and filters is shown in following Fig. for different filter order i.e. N=8, 16, 32. The ut off frequeny used in the frequeny range of SONAR (1kHz F 1kHz). So we used F 1 =1kHz, F2=1kHz and the sampling frequeny Fs=3kHz. The ombined Magnitude response and the group delay of the band pass filter using filter, shev1 filter, shev2 filter and filter for N=8, F 1 =1kHz, F 2 =1kHz and Fs=3kH is shown in Fig.3 and Fig The ombined Magnitude response and the group delay of the band pass filter using filter,shev1 filter,shev2 filter and filter for N=16, F 1 =1kHz,F 2 =1kHz and Fs=3kH is shown in Fig. and Fig hebyshev Fig. Comparison of magnitude response of different filters for N= ellipti elliptib Fig.3 Comparison of magnitude response of different filters for N= BUTTERWORTH CHEBYSHEV1 CHEBYSHEV2 ELLIPTIC Fig. 6 Comparison of Group delay of different filter for N=16 The ombined magnitude response and the group delay of the band pass, shev1, shev2 and filters for N=32, F 1 =1kHz,F 2 =1kHz and Fs=3kH is shown in Fig.7 and Fig. Comparison of Group delay of different filters for N=8

4 Group delay (in samples) International Journal of Emerging Tehnologies and Engineering (IJETE) ISSN: ICRTIET-21 Conferene Proeeding, 3 th -31 st August Fig ellipti Fig. 7 Comparison of magnitude response of different filters for N= ellipti Fig. 8 Comparison of Group delay of different filters for N=32. COMPARITIVE ANALYSIS Table. 1 Comparison of different filters with following parameters Order N=8 N=16 N=32 Protot ype-e filter Trans it-ion Width Stop Band Attenua tion-on Pass band Attenua tion-on Comparison of different filter tehniques in terms of filter parameters by varying the order an be observed from the different magnitude response of the band pass filter is shown in Table.1. The omparative analysis of transmission width, Pass band attenuation, Stop band attenuation with hanging the order for designing of different prototype filter is shown in terms of bar hart. Fig. shows the transmission width of different filter with different order. Fig. 1 shows the stop band attenuation of different filter with different order. Fig. 11 shows the pass band attenuation of different filter with different order. Fig. Transmission width (Hz) of different filter with different order

5 International Journal of Emerging Tehnologies and Engineering (IJETE) ISSN: ICRTIET-21 Conferene Proeeding, 3 th -31 st August Fig.1 Stop band attenuation (-db) of different filter with different Marh-27. order Fig.11 Pass band attenuation (-db) of different filter with different order. CONCLUSION From observation and omparison of different parameters of filters we analyzed that the response of filter is good at lower order in terms of its magnitude response and group delay. As we inrease the order the performane of ellipti filter is getting better ontinuously as ompared to other filters. From observation we an see that the response of hebyshev1 filter in terms of its stop band attenuation is more as ompared to others. As we inrease order the transition width of getting better and stop band attenuation and pass band attenuation is also getting better. Aknowledgement The authors would also like to thank Diretor, National Institute of Tehnial Teahers Training & Researh, Chandigarh, India and H.R. Institute of Tehnology, Ghaziabad for their onstant inspirations and support throughout this researh work. Referenes [1] Sanjit K. Mitra, Digital Signal Proessing, A Computer based Approah, Seond Edition, Page2-3. [2] S Salivahanan, A, Vallavaraj, C Gnanapriya, Digital Signal Proessing, Page38-381, 2. [3] John G. Proakis, Dimitris G. Manolakis, Digital Signal Proessing Priniples, Algorithms, and Appliations, Third Edition, Page682-68, 16. [] Ravi Kant Doneria, Prof. Laxmi Srivastava Comparative study of RF/Mirowave IIR Filters using matlab,volume 2,Issue-11,November 213,ISSN: ,Page [] Math works, Users Guide Filter Design Toolbox-, [6] Anju, Mamta Katiyar, Design of and shev1 Low pass Filter for Equalized, Volume 2, Issue-, May 212, Page2-28. AUTHORS Vanshikha Singh reeived the Bahelors of Tehnology degree in Eletronis and Communiation Engineering from Chandra Shekhar Azad University, Kanpur, India in 21.She is pursuing Master of Engineering degree in Eletronis and Communiation Engineering from National Institute of Tehnial Teahers Training & Researh, Punjab University, Chandigarh, India. Rajesh Mehra reeived the Bahelors of Tehnology degree in Eletronis and Communiation Engineering from National Institute of Tehnology, Jalandhar, India in 1, and the Masters of Engineering degree in Eletronis and Communiation Engineering from National Institute of Tehnial Teaher s Training & Researh, Punjab University, Chandigarh, India in 28. He is pursuing Dotor of Philosophy degree in Eletronis and Communiation Engineering from National Institute of Tehnial Teaher s Training & Researh, Punjab University, Chandigarh, India. He is an Assoiate Professor with the Department of Eletronis & Communiation Engineering,, National Institute of Tehnial Teaher s Training & Researh, Ministry of Human Resoure Development, Chandigarh, India. His urrent researh and teahing interests are in Signal, and Communiations Proessing, Very Large Sale Integration Design. He has authored more than 17 researh publiations inluding more than 1 in Journals. Mr. Mehra is member of IEEE and ISTE.