Performance Evaluation of Adaptive Noise Canceller Based on Multirate Filter Technique

Transcription

1 Performance Evaluation of Adaptive Noise Canceller Based on Multirate Filter Javaid Ahmad Sheikh Shabir Ahmad Parrah Jai Preet Kour Wazir G. Mohiuddin Bhat Abstract: This paper presents the structure of the Adaptive Noise Canceller (ANC) using the Least Mean Square (LMS) algorithm based on Multirate filters. The filtering idea with the ANC has been employed due to the fact that ANC is not able to filter out the distortion from the signal completely. Further the comparison between Cascaded Integrated Comb (CIC) Decimation filter and Adaptive filter has also been investigated. The performance in terms of Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR) have been performed it have been observed that Multirate filter reveals better results than Adaptive filter and the simulation results have been presented. Keywords: Adaptive Noise Canceller (ANC), Multirate filters, Adaptive filter, Least Mean Square algorithm (LMS), Cascaded Integrated Comb (CIC) Decimation filter. I. INTRODUCTION Noise or unwanted signal is presented in almost all signals. Noise can occur because of many disturbances. In some cases it is negligible; in other cases it eliminates the desired information from the signal. Actually, it is unwanted disruption that obstructs the desired information while keeping the source signal. The sources might comprise Speech, Music playing in different devices such in a mobile, laptops, or no sound at all. Noise can arise due to several reasons such as interference, interruption, and overlapping. To remove unwanted noise from the signal has been historically a driving force behind the enhancement of signal processing technology, and it continues to be a most important application for Digital Signal Processing Systems [3]. Adaptive Noise cancellation which is an effective method to remove the unwanted signal (noise) from the signal and leaves the desired signal. The adaptive noise canceller has two inputs: i) a primary signal corrupted by additive noise ii) a reference signal The adaptive filter processed the reference signal to minimize the power of the output signal. The design of adaptive filters has ability to adjust their own parameters automatically, and their design doesn t require prior or little knowledge of noise or signal. This minimization is attained by filtering the reference signal to estimate the noise in the primary signal and then subtracting it from the primary signal. Adaptive noise canceller is used with the adaptive filters and is employed for variety of applications like in Speech, Music, Electro-cardiac and Images [2]. Actually the Adaptive filter is a linear filter. Adaptive filter is employed in many applications due to the fact that it can change its parameters, as in some applications parameters of desired signal are not known in advance or changing but cannot eliminate the noise properly. Ideally, the output of the ANC is undistorted signal but practically contains some unwanted signal. Therefore, in order to remove noise completely and improve the performance of ANC, CIC decimation filter is used in our work. Systems that use different sampling rates at different stages are called the Multirate Systems. The Multirate techniques are used to convert the given sampling rate to the desired sampling rate and to provide different sampling rates through the system without destroying the signal components of interest. Multirate Systems in the processing of the digital signals have multiple sampling rates. The Multirate digital filters are used in many communication applications. They are also used to remove the undesirable signals such as noise and extract the valuable information signal. The work presented in this paper are able to remove the noise to the larger extend as have been verified by various parameters. Also, it is applicable for VLSI design due to low power requirement and less complexity. II. RELATED WORK Abishek Chaudary et_al (2014) [1] authors have implemented NLMS (Normalized Least Mean Square) algorithm rather than LMS. The authors have shown that with the increase in the scaling factor, the error signal increases and take much time to optimize error in accordance with the desired signal. The RES Publication 2012 Page 35

2 NLMS has faster convergence rate, more robust and results good performance due to its properties. The biggest disadvantage of NLMS algorithm is that it is unstable. In our work we have implemented ANC using LMS algorithm which is stable than NLMS. Vaibhav Purwar et_al (2013) [2] authors have not proposed any technique. In their work they have simply employed LMS adaptive filter only to automatically match the filter response. It only converges to the correct filter model. They have only employed the adaptive filter to remove noise from the signal. The authors have not employed any multirate technique to overcome the issue of computational complexity and slow convergence. Prasheel V. Suryawanshi et_al (2012) [3] the authors have proposed a new structure of ANC with multirate filters. In this proposed scheme the authors have shown work that the multirate filters overcome the computational complexity and slow convergence but only by calculating computational time and not in terms of PSNR, MSE and have not given any comparative study. III. ADAPTIVE NOISE CANCELLL ER to decrease an error. With the use of adaptive filter, noise reduces. ANC makes possible attainment to reject noise to a certain level. Assume that s, n, n1 and y are statistically stationary and have zero means. Assume that s is uncorrelated with n and n1, and suppose that n1 is correlated with n1. The output y is Squaring both sides The signal power minimize (1) Taking expectation both sides remains unaffected, filter is attuned to. Then, the minimum output power is (2) (4) (3) Adaptive noise cancellation is a technique of removing unwanted signal from the desired signal. The main purpose is to produce an estimate of noise and desired signal. The first adaptive noise cancelling system was designed at Stanford University in 1965 by two students. From the time 1965, ANC has been effectively useful to many problems, including other phases of electrocardiography, to eliminate the periodic interference, in telephone transmission lines of far areas to eliminate echoes, also in applications of the adaptive antennas. The basic model of ANC is shown in Figure 1. The signal is transmitted to a sensor that receives uncorrelated noise n and the signal x. Thus the combined signal x + n forms the primary input of the ANC. The second sensor receives noise n 1 which is correlated with the some noise while uncorrelated with the signal. It acts as a reference input of ANC. The noise n is then filtered to give an output z i.e., near replica of n. Then the output is subtracted from the primary input x+ n, to result the output, x+ n - z. It is necessary to have some prior knowledge of the input signal x or of the noises n and n 1 before designing the filter, or before it could adapt, to give the noise cancelling signal (z) [3]. The adaptive filter processes the reference input and adjusts its response automatically using LMS algorithm and produces error signal depending upon the filters output. Hence the adaptive filter that is capable of adjusting its impulse response Once the filter is adjusted to minimized, also minimized. Figure 1. Basic structure of Adaptive Noise Canceller (ANC) Adaptive filters are digital filters with an impulse response, or transfer-function that can be adjusted or changed over time to match desired characteristics. Disadvantages of adaptive filters. x + n' n' They assume that dynamic processes are linear. Merely gives a point estimate. z is y RES Publication 2012 Page 36

3 Can only handle processes with additive, uni-modal noise. A. Cascaded Integrated Comb Decimation Filter Cascaded Comb filter are formed by cascading comb filters and integrated and down sampler. The Cascaded Integrated Comb filter consists of equal number of integrator and filter pairs. Cascaded Comb filter is widely used as decimator and interpolator. Cascaded Comb filter allows efficient implementation with highly symmetric structure. CIC decimators due to simplicity are widely it doesn t use multiplication coefficients but uses addition and subtraction coefficients. The transfer function of composite CIC filter referenced to high sampling frequency is, Figure3. Actual signal correlated with noise (5) R = Decimation factor N = Number of stages. IV. PROPOSED TECHNIQUE The proposed work in Figure 4 involves ANC based on Multirate Filter. In a proposed technique an input signal is corrupted with Additive White Gaussian Noise (AWGN). Now the input signal corrupted with noise is transmitted through ANC uses LMS algorithm with step size µ= The step size µ controls the stability and the convergence speed. ANC does not remove the noise completely and is transmitted through filters one by one in order to check the performance of the filters. The output of both the filters are compared based on PSNR, MSE and computation time. The Multirate CIC Decimation Filter reveals good results than Adaptive Filter as shown in Table 1. Figure4. Output of Adaptive Filter Figure5. Output of CIC Filter V. SIMULATION Figure2. Actual signal correlated with noise In this paper the proposed technique has been verified on a deterministic signal. To evaluate the performance of ANC with Adaptive and CIC Filters, various performance parameters related to filters are calculated. The comparison of Adaptive Filter and CIC Filter are done on the basis of Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and RES Publication 2012 Page 37

4 Computational Time. The above quantities are calculated using the following formats: MSE = { }/N Where err is the error signal N is the size of original signal. PSNR = 10 log 10{max (input signal)/mse} Where, input signal is the original signal. Figure 2. The block diagram of proposed technique and Figure 3 shows input signal and input signal correlated with noise. The response of ANC with adaptive filter is shown in Figure 4 respectively. The output of ANC with (CIC) filter with R=8, M=1, N=4 is shown in Figure 5. Table1. Comparison of CIC Decimation filter and Adaptive Filter Parameters Adaptive Filter CIC DEcimation Filter with R=8,M=1,N =4 MSE PSNR Computation Time Parameter s ms ms Table2. Comparison of computation time with existing Proposed Existing Techniqe(Prah eel V.Suryawanshi et_al) [3] Figure7. Comparison between PSNR of Adaptive Filter and CIC Filter Input Deterministi Deterministic Signal c Figure8. Comparison between MSE of Adaptive Filter and CIC Filter Computati ms 4.2 ms on Time VI. CONCLUSION This work presents a new adaptive noise cancellation technique based on Multirate filters. The information signal in any system gets affected mainly due to noise and thus we chose the Adaptive noise canceller application. Noise is a random signal, in order to reduce its affect, the multirate filters are used and the filter coefficients will also change according to the signal. This work also investigates the comparison performance of CIC Decimation filter and Adaptive filter. Simulation results show the performance of CIC Decimation filter in terms of MSE, PSNR, and Computational Time. CIC Decimation filters can also be used to overcome the problem of Computational complexity and slow convergence rate problem. From the above observations it is concluded that proposed techniques can serve as a back bone for morden high speed multirate Figure6. Comparison between Proposed and Existing RES Publication 2012 Page Adaptive Filter MSE Adaptive Filter CIC Filter CIC Filter

5 system such as software Defined Radios (SDR), Biomedical instrument, Cognitive radios and many more. VII. REFERENCES [1]. Abishek Chaudary, Amit Barnawal, Anushree Gupta, Deepti Chaudhar, Analysis of Noise Signal Cancellation using Adaptive Algorithms, International Journal of Engineering Research and General Science 2014; Vol 2 Issue 6, [2]. Gyanendra Singh, Vaibhav Purwar, Kiran Savita,Shivkumar Yadav, Design of Adaptive Noise Canceller using LMS Algorithm, International Journal of Advanced Technology & Engineering Research (IJATER) 2013; ISSN No: Vol 3 Issue 3. [3]. Prasheel V.Suryawanshi, Kaplil prasad Mahapatro, Vardhman J.Sheth, Adaptive Noise cancellation using Multirate s, International Journal of Engineering Research and Development 2012; ISSN: X, Vol 1 Issue 7: [4]. Javaid A Sheikh, Jai Preet Kour Wazir, Shabir A. Parah G.M.Bhat, On the Study and Performance Evaluation of Multirate Filter, Commune 2015 International conference on Advances in Computer, Communication and Electronics Engineering; [5]. Javaid A Sheikh, Jai Preet Kour Wazir, Shabir A. Parah G.M.Bhat, Design and Realisation of Multirate Signal Processor for Enhancement Performance of Digital Receivers, 12th IEEE India International Conference (INDICON) on Electronics, Energy, Environment, Communication, Computers, Control (E3-C3), December, 2015, Jamia Millia Islamia, New Delhi. INDICON 2015 IEEE India Council. [6]. Riggi Aquino and Jacob Lincoln, Hardware and Software Study of Active Noise Cancellation. Project Report, California Polytechnic State University [7]. Ljiljana Milic, Multirate Filtering for Digital Signal Processing: MATLAB Applications.Hershey, PA: Information Science Reference Jan [8]. Richard G.LYONS. Understanding Digital Signal Processing, Prentice Hall [9]. Gordana Jovanovic Dolecek and M.Laddomada, An improved class of multiplierless decimation filters: Analysis and design Elsevier June 2013; [10]. Gordana Jovanovic Dolecek and Afono Fernandez-Vazquez. Novel droop-compensated comb decimation filter with Improved alias rejections, SciVerse Science Direct October 2012; [11]. Paulo S.R. Diniz. Adaptive Filtering Algorithm and Practical Implementation, Springer Publication 2008; AUTHOR S BIOGRAPHY Javaid A. Sheikh has completed his M.Sc., M. Phil and Ph. D in Electronics from, Srinagar in the year 2004, 2008 and 2012 respectively in the field of communications and Signal Processing. He is working as Assistant Professor in the department of Electronics and I. T, Srinagar. His field of interest are Wireless Communications, design and development of efficient MIMO- OFDM based wireless communication techniques, Spread Spectrum modulation, Digital Signal Processing, Electromagnetics and Speech Processing and compression techniques. Jai Preet Kour Wazir has completed her M.Sc and M.Phil in Electronics from, Srinagar in the year 2012 and 2016 respectively in the field of Communication and Multirate s. She is presently pursuing her Ph.D. in the same field from the. Shabir A. Parah has completed his M. Sc, M. Phil and Ph. D in Electronics from University of Kashmir, Srinagar in the year 2004, 2010 and 2013 respectively in the field of Signal processing and Data Hiding. He is working as Assistant Professor in the department of Electronics and I. T University of Kashmir, Srinagar. His field of interest are Signal Processing, Embedded Systems, Secure Communication and Digital design and Watermarking. He has published more than 60 research papers in International and National journals and conference proceedings. Prof. G. Mohiuddin Bhat obtained his M.Sc. (Electronics) from the, Srinagar (India) in 1987, M.Tech. (Electronics) from Aligarh Muslim University (AMU), Aligarh (India) in 1993 and Ph.D. Electronics Engg. From AMU, Aligarh, (India) in He has served as Assistant Professor, Associate professor, Professor and Director in University Science Instrumentation Centre (USIC),. Presently Prof. G. M. Bhat is Dean Faculty of Applied Science and Technology,. He has worked in the area of Mobile Radio Communication, Spread Spectrum Communication and Neural Networks and has guided many research degrees leading to the award of M.Phil and Ph.D. RES Publication 2012 Page 39