Active Noise Cancellation Using Adaptive Filter Algorithms

International Journal of Engineering Research & Technology (IJERT)ISSN: 2278-0181Vol. 7 Issue 02, February-2018Published by :http://www.ijert.orgActive Noise Cancellation using Adaptive FilterAlgorithmsB A SujathakumariRavi S BarkiAssociate ProfessorSri Jayachamarajendra College of EngineeringJSS Science &Technology University, Mysuru, India1st Sem M TechSri Jayachamarajendra College of EngineeringJSS Science &Technology University, Mysuru, IndiaLokesh A RPavithra C M1st Sem M TechSri Jayachamarajendra College of EngineeringJSS Science &Technology University, Mysuru, India1st Sem M TechSri Jayachamarajendra College of EngineeringJSS Science &Technology University, Mysuru, IndiaAbstract:- This paper describes a speech enhancement systembased on TMS320C6713 digital signal processor (DSP) for activereal-time noise cancellation. In conventional FIR and IIR digitalfilters, it is assumed that the process parameters to determine thefilter characteristics are known. They may vary with time, butthe nature of the variation is assumed to be known. In manypractical problems, there may be a large uncertainty in someparameters because of inadequate prior test data about theprocess. Some parameters might be expected to change with time,but the exact nature of the change is not predictable. In suchcases it is highly desirable to design the filter to be self-learning sothat it can adapt itself to the situation at hand. In this paper the focus is on enhancement of speech audiosignal from noise and will compare Least Mean-Square (LMS),Normalized Least Mean-Square (NLMS) and Recursive LeastSquare (RLS) algorithms using DSP processor with codecomposer studio (CCS)II NOISE CANCELLATION USING ADAPTIVE FILTERThe Single channel adaptive feedback Active NoiseCancellation (ANC) system works by processing the acousticalnoise (the “target” noise) that would like to reduce, byproducing an anti-noise which cancels out the noise componentby the method of adaptive filtering. Therefore, the main aim ofANC system is to reduce the noise component from the signalof interest. Figure 1 shows schematic diagram of a singlechannel feedback ANC systemKeywords: Adaptive noise cancellation (ANC), LMS algorithm,NLMS algorithm, RLS algorithm, adaptive filterI. INTRODUCTIONAdaptive filters are best used in cases where signal conditionsor system parameters are slowly changing and the filter is tobe adjusted to compensate for this change. A very simple butpowerful filter is called the linear adaptive combiner, which isnothing more than an adjustable FIR filter.The LMS criterion is a search algorithm that can be used toprovide the strategy for adjusting the filter coefficients.In conventional FIR and IIR digital filters, it is assumed thatthe process parameters to determine the filter characteristicsare known. They may vary with time, but the nature of thevariation is assumed to be known. In many practical problems,the coefficients of an adaptive filter are adjusted tocompensate for changes in input signal, output signal orsystem parameters. Instead of being rigid, an adaptive systemcan learn the signal characteristics and track slow changes. Anadaptive filter can be very useful when there is uncertaintyabout the characteristics of a signal or when thesecharacteristics change.Fig1: Single channel feedback active noise cancellationThe anti-noise waveform is similar to that of the target noise,except its phase is reversed by 180 degrees. If these waveformsare combined together. It will result in a much weaker residualwaveform (this waveform should be of zero amplitude if theanti-noise matches perfectly with the target noise).The residual waveform is what is picked up by the microphoneas shown in figure2. The feedback ANC system produces ananti-noise by predicting the incoming target noiseFig 2: Description of active noise controlIJERTV7IS020020www.ijert.org(This work is licensed under a Creative Commons Attribution 4.0 International License.)15

Published by :http://www.ijert.orgInternational Journal of Engineering Research & Technology (IJERT)ISSN: 2278-0181Vol. 7 Issue 02, February-2018NOISE CANCELLATION USING ADAPTIVE FILTERALGORITHMS(i) LEAST MEAN SQUARE (LMS) ALGORITHMIn the LMS algorithm, the coefficients are adjusted fromsample to sample in such a way as to minimize the MeanSquare Error (MSE).The LMS is based on the steepest descentalgorithm where the weight vector is updated from sample tosample as followsIII.Wk 1 Wk - µk(1)Where Wk and µk are the weight and the true gradient vectors,respectively at the kth sampling instant. µ controls the stabilityand rate of convergence.1.Initially, set each weight to an arbitrary fixed value as 0.For each subsequent sampling instant, k, carryout stepsfrom 2 to 4 below.2.Compute filter outputn’k Σw (i) x k-i3.Compute the error estimateek yk – n’k4.(2)(3)Update the next filter weightswk 1(i) wk(i) 2µekx k-i(4)Fig 3: Flowchart of LMS algorithmBecause of its simplicity, the LMS algorithm is one of thepopular adaptive algorithm. However, the LMS algorithm isvery slow and data dependent convergence behaviour.One of the primary disadvantages of the LMS algorithm ishaving a fixed step size parameter for every iteration. Thisrequires an understanding of the statistics of the input signalprior to commencing the adaptive filtering operation. Inpractice this is rarely achievable.Implementation of Least Mean Square Algorithm for noisereductionInitially, the weight parameter (w) and also the loop variable(i) are set to zero. Then obtained input signals from themicrophone. In the next step the filter output is calculated andis further used to compute the error estimate signal.Then the filter weights are updated by fixing the step sizevalue (µ) to be a constant. This procedure is repeated until theloop parameter becomes equal to the buffer size. Thisimplementation is depicted in figure 3.Stability OF LMS: Condition for stability is: 0 µ 2 (input signalpower)Larger values for step sizeIncreases adaptation rate (faster adaptation)Increases residual mean-squared error(II) RECURSIVE LEAST SQUARE (RLS) ALGORITHMRLS is relatively complex algorithm as compared to LMS andNLMS algorithm. Also performance of RLS in terms ofconvergence and Mean Square Error (MSE) is better whencompare to LMS and NLMS.The Recursive least square (RLS) adaptive filter is analgorithm which recursively determines the filter coefficientsthat reduces a weighted linear least squares cost functionrelating to the input signals. The RLS algorithms are knownfor their excellent performance when working in time varyingenvironments but at the cost of an increased computationalcomplexity along with stability problems.IJERTV7IS020020www.ijert.org(This work is licensed under a Creative Commons Attribution 4.0 International License.)16

Published by :http://www.ijert.org1.International Journal of Engineering Research & Technology (IJERT)ISSN: 2278-0181Vol. 7 Issue 02, February-2018RLS adaptation algorithm with input signals y(n) and x(n)is given below. Initial values for RLS algorithm is givenbyP (n) δIP (0) (δ 1) Iw (0) wI2.(5)(6)(7)Inverse correlation matrix update is calculated usingP(n) λ 1[P(n 1) (λ 1)k(n)yT(P(n 1)](8)(iii) NORMALIZED LMS ALGORITHM (NLMS)The NLMS algorithm is equally simple, but more robustvariant of the LMS algorithm and exhibits a better balancebetween simplicity and performance compared to LMSalgorithm. Due to its better properties the NLMS has beenlargely used in real-time applications. The normalized leastmean square algorithm (NLMS) is an extension of the LMSalgorithm which bypasses this issue by computing maximumstep size value. Step size value is computed by using thefollowing formula.Step size 1/dot product (input vector, input vector)This step size is proportional to the inverse of the totalexpected energy of the instantaneous values of the coefficientsof the input vector. NLMS algorithm is having the advantageover the LMS algorithm in case of Mean square error andAverage attenuationIMPLEMENTATION OF NLMS ALGORITHM1.(11)Filter coefficient adaptation is given byw(n) w(n 1) k(n)e(n)5.P’k 1 (1-β) P’k Nβx2k-iWhere, β: smoothing parameterError signal equation is given bye(n) x(n) wT(n 1)y(n)4.Power estimation is upgraded by following equation,For m 1,2, filter gain update vector is given byk(n) λ 1P(n 1)y(n) (λ 1)yT(n)P(n 1)y(n)3.2.The NLMS algorithm is written as followsWk 1 Wk 2µekXk(9)µk α / NP’Where, µk : time varying step size factor normalised by NN : filter lengthα : constant ( 0 α 2 )P’k: power estimate of xkIJERTV7IS020020Fig 4: Flowchart of NLMS algorithmInitially, the weight parameter (w) and the loop variable (i) areset to zero. The obtained input signals from the microphone. Inthe next step the filter output is calculated and is further usedto calculate the error estimate signal. The filter weights areupdated depending on the calculated step size value (µ). Thisprocedure is repeated until the loop parameter becomes equalto the buffer size. This implementation is depicted in figure 4.REAL-TIME IMPLEMENTATION THROUGH CODECOMPOSER STUDIOCCS provides an IDE to incorporate the software tools. CCSincludes tools for code generation, such as a C compiler, anassembler, and a linker. It has graphical capabilities and alsosupports real time debugging and provides an easy-to-usesoftware tool to build and debug programs. The C compilercompiles a C source program with extension .c to produce anassembly source file with extension.asm. The assemblerassembles an .asm source file to produce a machine languageobject file with extension.obj. The linker combines object filesand also objects libraries as input to produce an executable filewith extension.outReal-time analysis is performed using real-time data exchange(RTDX). RTDX allows for data exchange between the hostPC and the target DSK, along with analysis in real timewithout stopping the target. Key statistics and performance canbe monitored in real time. Through the joint team action group(JTAG), communication with on-chip emulation supportoccurs to control and monitor program execution. The C6713DSK board includes a JTAG interface through the USB port.www.ijert.org(This work is licensed under a Creative Commons Attribution 4.0 International License.)17

Published by :http://www.ijert.orgInternational Journal of Engineering Research & Technology (IJERT)ISSN: 2278-0181Vol. 7 Issue 02, February-2018IMPLEMENTATION USING TMS320C6713 PROCESSORIN REAL TIMEAlgorithm implementation is done on TMS320C6713processor using Code Composer Studio (CCS) software whichgenerates the executable code of the algorithm and burns thesame to the DSK chip. The features of TMS320C6713Processor are as explained below. VLIW core – is a processor architecture that allows 8instructions to be issued in a single clock.192KB of high speed internal memory for maximummemory performance.64KB L2 cache/RAM four 16KB blocks of internalRAM that can be configured as RAM/cache.4KB separate caches for program and data.On – chip PLL generates processor clock rate fromslow external clock reference.Two timers generate periodic times events as afunction of processor clock. Used by DSP/BIOS tocreate time slices for multi – tasking.Enhanced DMA (EDMA) controller that allows highspeed data transfer without intervention from DSP.IV.Fig7: Output: Mean Square Error for NLMSFig8: Mean Square Error for RLSEXPERIMENTAL RESULTS Output: Mean Square Error For LMS.Fig 9: Actual signal Fig5: Mean Square Error for LMS Fig 10: Input and noise signalFig6: Variation Of MSE With Respect To ΜFig 10: LMS filter outputIJERTV7IS020020www.ijert.org(This work is licensed under a Creative Commons Attribution 4.0 International License.)18