The document discusses various methods for solving difference equations in discrete time domain including iterative, filter command, convolution, and graphical methods. It provides examples of applying each method to find impulse response, zero input response, and total response. Exercises are given to apply the methods to specific difference equations and analyze the nature, stability, time constant, and response characteristics.
The document discusses various methods for solving difference equations in discrete time domain including iterative, filter command, convolution, and graphical methods. It provides examples of applying each method to find impulse response, zero input response, and total response. Exercises are given to apply the methods to specific difference equations and analyze the nature, stability, time constant, and response characteristics.
The document discusses various methods for solving difference equations in discrete time domain including iterative, filter command, convolution, and graphical methods. It provides examples of applying each method to find impulse response, zero input response, and total response. Exercises are given to apply the methods to specific difference equations and analyze the nature, stability, time constant, and response characteristics.
The document discusses various methods for solving difference equations in discrete time domain including iterative, filter command, convolution, and graphical methods. It provides examples of applying each method to find impulse response, zero input response, and total response. Exercises are given to apply the methods to specific difference equations and analyze the nature, stability, time constant, and response characteristics.
Method 1: Solving Difference equation using Iterative procedure :
Finding Total Response:
Ex: Given the difference equation, y[n+2] – y[n+1] + 0.24y[n] = x[n+2] – 2x[n+1] and with initial conditions : y[-1] = 2 ; y[-2] = 1, with x[n] = n (starting at n = 0) Following is the MATLAB code: n = [-2:10]'; % defining the discrete time range for the solution y = [1 ; 2 ; zeros(length(n) -2 , 1)]; % predefining the output array with first two elements are initial conditions and subsequent elements are zeros x=[ 0; 0 ; n(3 : end) ]; % defining the input array with first two entries corresponds to the time instants -2 and -1 for k = 1 : length(n) -2 % Iterative computation starts here y(k+2) = y(k+1) - 0.24*y(k) + x(k+2) - 2*x(k+1); end; % NOTE : when k = 1, y(3) value is computed indicating the first member of the solution. Since, y(1) and y(2) have already assigned with initial conditions. stem(n , y , 'k') % sketch of the solution including the initial conditions. O/P -----> y = [1.0000 2.0000 1.7600 2.2800 1.8576 0.3104 -2.1354 - 5.2099 - 8.6974 - 12.4470 -16.3597 -20.3724 -24.4461] ▪ For finding the zero input response, do not consider the feed forward coefficients (terms corresponding to the coefficients of x[n] and its advance operations), in the difference equation. OR define the input array as zeroes ▪ For finding the zero state response, define the initial conditions as zeroes ▪ For finding the impulse response, define the input array as impulse and initial conditions as zeroes. i.e. x=[ 0; 0 ; 1 ; zeros(length(n)-3 , 1)];
Method 2: Solving Difference equation using “filter” command in MATLAB:
MATLAB’s filter command provides an efficient way to evaluate the system response to a constant coefficient (time-invariant) linear difference equation represented in delay form as 𝑵 𝑵
∑ 𝒂𝒌 𝒚[𝒏 − 𝒌] = ∑ 𝒃𝒌 𝒙[𝒏 − 𝒌] 𝒌=𝟎 𝒌=𝟎 Filter requires three input arguments: a length N+1 vector of feed forward coefficients [bo,b1,b2,…bN]
1|Page a length N+1 vector of feedback coefficients [ao,a1,a2,…aN] and an input vector. Optional 4th argument: initial state of the system should be appropriately inserted
Finding Impulse Response:
For the same above example, following is the MATLAB code: N = 2; b = [1 -2 0]; a = [1 -1 0.24]; %defining difference equation n = [0 : 30]' ; delta = inline('n == 0','n'); %defining impulse function h = filter(b,a,delta(n)); stem(n,h) Result: h = {1 -1 -1.24 -1 -0.702 -0.462 -0.294 -0.183 -0.1123 .....} NOTE : ‘filter’ function gives first N (i.e. order of system) values of the solution as auxiliary conditions, which is generally counted from running variable n = 0 to N-1. Here, first two values {1 and -1} in above result are considered as auxiliary conditions of the solution.
Finding Zero –State Response:
For the same above example, following is the MATLAB code: N = 2; b= [1 -2 0]; a= [1 -1 0.24]; %defining difference equation n=[0:30]'; x = inline(' n ',' n ') %defining input function ys= filter(b,a,x(n)); stem(n,ys) Result : ys = {0, 1, 1, -0.24, -2.48, -5.42, -8.82, -12.5258 . .......} NOTE : “filter” function can’t be used in MATLAB to find the ZIR of the difference equation, as the initial conditions can’t be accommodated directly. One may use the iterative method discussed above and it may be added to ZSR to get the total response.
Method 3 : Convolution of two signals using ‘conv’ command
Case 1): Convolution of two finite-duration discrete time signals is accomplished by using the conv command. “Conv” command does not give the output with the region of support in the time axis. Eg: 1) (u[n] - u[n-4) * (u[n] - u[n-4) >> conv( [1 1 1 1] , [1 1 1 1] ) O/P ----> 1 2 3 4 3 2 1 Regions of support in time axis is (0 ≤ n ≤ 6) (7 elements = 4 + 4 - 1) 2) (u[n+4] - u[n) * (u[n] - u[n-4) >> conv( [1 1 1 1] , [1 1 1 1] )
2|Page O/P ----> 1 2 3 4 3 2 1 Regions of support in time axis is (- 4 ≤ n ≤ 2) (graphically supported) The region of support for thus obtained output may be comprehended by the width property.
Case 2): In general, the conv command should be carefully used to convolve infinite – duration signals. Conv command can correctly compute a portion of such infinite duration signals.
By passing the first N samples of each signal, conv command returns a length (2N - 1) sequence. The first N samples are computed correctly; the remaining N-1 samples are wrong results.
For the difference equation mentioned in the above example, the impulse response (calculated analytically) of the system is h[n] = {-7 (0.6)n + 8 (0.4)n } u[n], hence by defining the inline function calling it with n = [0 : 30] ' as: >> h = inline('(-7*0.6.^n+8*0.4.^n).*(n>=0)','n'); >> x = inline(' n ',' n '); >> ytc= conv(h(n),x(n)); >> stem ( [0:60] , ytc ,'k') Compare the result with earlier method.
In the result, 0 ≤ n ≤ 30, output samples are correctly computed. The remaining 31 ≤ n ≤ 61 output sample values are incorrectly computed. Theoretically, the output envelope should continue to grow as n tends to infinity, not decay. If “conv” command is used to convolve infinite – duration signals, normally incorrect values should be omitted and are not used for display as: >> stem(n,ytc(0:30),'k') will be identical to the earlier result.
Method 4 : Convolution of two signals : (using Graphical Method)
Try to modify the MATLAB code to convolve two discrete time signals graphically and use it. (Refer the MATLAB E code discussed in Module 3 part 1). Hint: one may use “stem” instead of plot. Also, stem sketches only one signal in one axis. One may use many subplots to achieve this. And Increment considered is 1. In this case, irrespective of the signal duration, finite or infinite, one can get the output results with the regions of support in time axis. This method does give the output with the region of support in the time axis.
3|Page Exercises
For all the following exercises of A) to E), comment (where ever applicable) on the Nature (characteristic polynomial, characteristic roots, characteristic modes), stability, and find the time constant (in terms of sample count) of the system. Sketch the output responses and comment on the nature of the responses (belongs to which class of signal model). Comment on the cut-off frequency and filtering action (by suitably changing the forcing function frequency). {EXPECTATION IS THAT HAVING UNDERSTOOD THE SAME IN CONTINUOUS_TIME, SAME CAN BE INTERPRETED HERE WITH MINOR DIFFERENCE}
A) Solve the following difference equations, to find the Impulse Response, Zero –Input Response, Zero–State Response and Total Response. Try to get the results iteratively (20 samples) and also using filter (as applicable) command. 1) y[n+2] –0.6 y[n+1] + 0.16y[n] = 0 with y[-1] = -25 ; y[-2] = 0 2) y[n+2] +3 y[n+1] + 2y[n] = x[n+2] +3x[n+1] +3x[n] with x[n] = (3)nu[n], y[-1] = 3 ; y[-2] = 2 3) y[n] +2 y[n-1] + y[n-2] = 2x[n] -x[n-1] with x[n] = (3)-nu[n], y[-1] = 2 ; y[-2] = 3 4) y[n+3] –1.6 y[n+2] + 0.8y[n+1] +0.02y[n] = x[n] with x[n] = u[n], y[-1] = 1 ; y[-2] = 2 and y[-3] =3
B) Consider the discrete time system y[n] + y[n-1] +0.25 y[n-2] = 3 x[n-8]. Find the zero input response, if y[-1] = 1 ; y[-2] = 1
C) Find the unit impulse response h[n] and hence find ZSR for the systems specified by the following difference equations: Try to get the results using graphical convolution (Method 4) AND using “conv” (method 3) functions and compare. 1) y[n+1] + 2y[n] =x[n] with x[n] = 2n u[n-1] 2) y[n] + 2y[n-1] =x[n] with x[n] = 3n+2u[n+1] 3) (E2-6E+9)y[n] = Ex[n] with x[n] = (3)-nu[n] 4) y[n] - 6 y[n-1] + 25y[n-2] = 2x[n] -4x[n-1] with x[n] = n u[n]
D) Find the zero state response of the LTID system, whose following unit impulse response h[n] and input x[n]. 1) h[n] = (-2)nu[n-1] ; x[n] = e-nu[n+1] 2) h[n] = ½{δ[n-2]-(-2)n+1 }u[n-3] ; x[n] = 3n-1u[n+2] 3) h[n] = {(2)n-2 +3(-5)n+2 }u[n-1] ; x[n] = 3n+2u[n+1] 4) h[n] = 3(n-2)(2)n-3 u[n-4] ; x[n] = 3-n+2 u[n+3] 5) h[n] = (3)n cos n − 0.5 u[n] ; x[n] = 2n u[n-1] 3 E) Use classical method, to solve the equations given in exercise A), and compare the results with the iterative methods.
