113_1_midterm_08SP_A

113_1_midterm_08SP_A - EE113: Digital Signal Processing...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: EE113: Digital Signal Processing Solution to Midterm Exam (A) Spring 2008 1. Consider the following two systems y1 (n) = 1 4 n x(k ), k=n−3 y2 (n) = x 0 n 3 n = 3k, k integer n = 3k, k integer. (a) Determine whether the systems are linear, time-invariant, relaxed, BIBO stable, and causal. Justify your answer to receive full credit. Solution: Properties Relaxed Linear Time-Invariant BIBO Stable Causal y1 (n) Yes Yes Yes Yes Yes y2 (n) No Yes No Yes No • System y1 (n): n 1 1 · y1 (n) = 4 k=n−3 x(k ) = 4 [x(n − 3) + x(n − 2) + x(n − 1) + x(n)] can be considered as a relaxed constant coefficient difference equation. It is relaxed, linear, time-invariant, BIBO stable, and causal. • System y2 (n): · Not relaxed: Counter example: if the first non-zero sample of x(n) is at n = −1, then y2 (−3) will be non-zero, which is before −1. · Linear: Let two outputs be a y2 (n) = xa 0 n 3 n = 3k, k integer , n = 3k, k integer. and b y2 (n) = xb 0 n 3 n = 3k, k integer , n = 3k, k integer. for an input xa (n) and xb (n), respectively. The output y (n) for an input Axa (n)+ Bxb (n) with some constants A and B is given by i) if n = 3k , k integer, y (n) = Axa n n + Bxb 3 3 a b = Ay2 (n) + By2 (n). ii) if n = 3k , k integer, y (n) = 0. Based on i) and ii), y (n) for an input Axa (n) + Bxb (n) is a b y (n) = Ay2 (n) + By2 (n), which implies that system y2 (n) is linear. · BIBO Stable: For a bounded sequence |x(n)| ≤ M < ∞ with a finite positive number M , |y2 (n)| = |x n | ≤ M < ∞ n = 3k, k integer 3 , 0<∞ n = 3k, k integer. implying that |y2 (n)| is bounded output. · Not Time-Invariant: Consider inputs δ (n) and δ (n − 1). The outputs are δ (n) and δ (n − 3). This shows that y2 (n) is not time-invariant. · Not causal: For example, y2 (−3) depends on x(−1), which is on the future of y2 (−3). (b) Given x(n) = nu(n), compute and plot y1 (n) and y2 (n) for 0 ≤ n ≤ 5. Solution: n y1 (n) y2 (n) 0 0 0 1 1/4 0 2 3/4 0 3 6/4 = 3/2 1 4 10/4 = 5/2 0 5 14/4 = 7/2 0 The corresponding plots are shown in Figure 1. (c) Find the z -transform of y1 (n) and y2 (n) in terms of the z -transform of x(n). Solution: z -transform for y1 (n) = 1 4 n k=n−3 1 x(k ) = 4 [x(n − 3) + x(n − 2) + x(n − 1) + x(n)] gives Y1 (z ) = 1 −3 [z X (z ) + z −2 X (z ) + z −1 X (z ) + X (z )] 4 z −3 + z −2 + z −1 + 1 X (z ). = 4 2 y (n) 1 y (n) 2 3.5 1 0.9 3 0.8 2.5 0.7 0.6 0.5 1.5 0.4 0.3 0.2 0.5 0.1 0 0 2 n 4 6 0 0 2 n 4 6 2 1 Figure 1: Plot for y1 (n) and y2 (n). z -transform for y2 (n) is given by ∞ Y2 (z ) = n=−∞ ∞ y2 (n)z −n x n=−∞ ∞ = = k=−∞ ∞ n −n z 3 x(k )z −3k x(k )(z 3 )−k k=−∞ 3 = = X (z ). 3 2. A causal system is described by the following block diagram. x(n) −1 6 z −1 y (n) 1 1 3 z −1 (a) Determine the constant coefficient difference equation that describes the system, and find its impulse response. Solution: Interchanging the order of the two systems in cascade, we get 1 1 y (n) = − y (n − 1) + y (n − 2) + x(n) + x(n − 1). 6 3 The impulse response can be found by computing the zero-state response to x(n) = δ (n). We first find the homogeneous solution: 1 1 λ2 + λ − = 0 6 3 and hence, yh (n) = C1 1 2 n ⇒ λ1 = 2 3 1 , 2 n λ2 = − 2 3 + C2 − , for all n. Letting x(n) = δ (n) and assuming that the system is relaxed, we get y (0) = 1 and y (1) = − 1 y (0) + 1 = 5 . From the homogeneous solution we get 6 6 y (0) = C1 + C2 = 1 1 2 5 y (1) = C1 − C2 = 2 3 6 and solving for C1 and C2 yields C1 = h(n) = 9 7 and C2 = − 2 , and consequently, 7 1 2 n 9 7 − 2 7 − 2 3 n u(n). (b) Given x(n) = n2n u(−n), find the output of the system using the z -transform. Solution: The output y (n) when x(n) = n2n u(−n) can be found by computing the inverse z -transform of Y (z ) = H (z )X (z ). The z -transform of h(n) is readily found from the impulse response: H (z ) = 1 + z −1 , 1+ − 1 z −2 3 1 −1 6z ROC: |z | > 2 . 3 4 The z -transform of x(n) is given by X (z ) = Z{x(n)} = −z = −z d dz ∞ d d Z{2n u(−n)} = −z dz dz n 0 2 n z −n n=−∞ n=0 z 2 = −z d 1 2z , =− 1 dz 1 − 2 z (2 − z )2 ROC: |z | < 2. Consequently, Y (z ) is given by Y (z ) = H (z )X (z ) = −2z 3 − 2z 2 , (z − 1 )(z + 2 )(z − 2)2 2 3 ROC: 2 < |z | < 2. 3 Using partial fraction expansion, we get A Y (z ) B = 1+ z z−2 z+ where A= B= C= z− z+ 1 2 2 3 Y (z ) z Y (z ) z =− 1 z= 2 2 3 + D C + z − 2 (z − 2)2 4 7 3 56 =− z =− 2 3 d Y (z ) 5 (z − 2)2 = dz z 8 z =2 2 Y (z ) = −3. D = (z − 2) z z=2 Thus, Y (z ) = − 4z 7z− − 2 3 1 2 3z 56 z + 2 3 + 5z 3 2z − 8 z − 2 2 (z − 2)2 n and since the ROC of Y (z ) is given by z z− < |z | < 2, we get ↔ ↔ ↔ ↔ 1 2 − 2 3 u(n) n 1 2 z 2 z+3 z z−2 2z (z − 2)2 and consequently, y (n) = − 4 7 1 2 n u(n) −2n u(−n − 1) −n2n u(−n − 1) − 3 56 − 2 3 n 53 u(n) + − + n 2n u(−n − 1). 82 Alternatively, partial fraction expansion on Y (z ) yields y (n) = − 2 7 1 2 n−1 + 1 28 − 2 3 n−1 5 u(n − 1) + − + 3n 2n−1 u(−n). 4 5 3. The difference equation of a relaxed system is: y (n) + 0.5y (n − 1) − 0.14y (n − 2) = x(n) (a) Find a closed form for the impulse response h(n) (i.e., the zero state system output when the input is an impulse). Solution: By setting x(n) = δ (n), we have h(n) + 0.5h(n − 1) − 0.14h(n − 2) = δ (n) = 1, n = 0 0, n = 0. Hence, for n ≥ 1, h(n) can be found by solving homogeneous difference equation h(n) + 0.5h(n − 1) − 0.14h(n − 2) = 0. The characteristic polynomial for (1) can be solved by λ+ thus, the modes are λ1 = − Hence, h(n) = C1 − 7 10 7 10 λ− 1 5 = 0, (1) 7 , 10 n λ2 = 1 5 1 . 5 n + C2 , n ≥ 0. Since the system is relaxed, y (−1) = h(−1) = 0 and h(0) = δ (0) = 1, which gives C1 = Therefore, h(n) = 7 9 − 7 10 7 , 9 n C2 = 2 . 9 1 5 n + 2 9 u(n). (b) If the input to the system is x(n) = 2δ (n) + δ (n − 2), what is the output? Solution: The output y (n) of the system can be expressed as the convolution of x(n) and h(n), i.e., y (n) = x(n) ∗ h(n) = [2δ (n) + δ (n − 2)] ∗ h(n) = 2h(n) + h(n − 2). 6 Hence, using h(n) in part (a) gives us i) n ≥ 2, y (n) = 2h(n) + h(n − 2) =2 7 9 − 7 10 n + 7 10 +6 2 9 −2 1 5 n u(n) + − 7 10 n 7 9 − 7 10 n−2 + 1 5 2 9 −2 1 5 1 5 n−2 u(n − 2) n = 2· = ii) 0 ≤ n ≤ 1, 22 7 77 +· 99 − 7 10 n + 2· 22 +· 99 1 5 n . y (n) = 2h(n) =2 = Therefore, y (n) = 0, 14 9 22 7 7 9 14 9 7 21 u(n) + 10 95 n n 7 41 − + . 10 95 − n n n < 0, 7n − 10 + 4 9 7n − 10 + 6 1n , 5 1n , 5 n = 0, 1, n ≥ 2. (c) For what values of α is g (n) = αn h(n) a finite energy sequence? Solution: The energy Eg of g (n) is given by ∞ Eg = n=−∞ ∞ |g (n)|2 |αn h(n)|2 n=−∞ ∞ = = n=0 ∞ αn 7 9 7 9 − − 7 10 n n + + 2 9 ∞ 2 9 1 α 5 1 5 n 2 = n=0 ∞ 7 α 10 n2 = 49 81 n=0 49 2 α 100 n + 4 81 n=0 12 α 25 n + 28 81 n=0 ∞ − 72 α 50 n . Thus, for the energy to be finite, 49 2 α < 1, 100 12 72 α < 1, and α < 1, 25 50 7 or equivalently, |α| < Hence, α should satisfy |α| < 10 7. 10 , 7 |α| < 5, and |α| < 50 . 7 8 ...
View Full Document

Ask a homework question - tutors are online