corrected
EE 341: Discrete Linear Systems Department of Electrical Engineering U NIVERSITY OF WASHINGTON HW 1.1 The following graph is of signal x[n] SOLUTION:Problem Set 1 Due Oct 9 in class Oct 2, 2009
Figure 1: Problem 1.1 Assuming that the signal is zero except as shown, plot the following: (a) 3x[n + 1]
Figure 2:
(b) x[2n] 1
Figure 3: (c) x[n] + 2
Figure 4:
HW 1.2 Sometimes signals can be decomposed into combinations of simple unit sequences (function u[n]), such as this one:
y [n] = 2u[n 2] 2u[n 7] 2u[n] + 2u[4 n] Sketch y [n] and the following signals.
Figure 5: y [n] (a) 2 3y [n]
Figure 6:
(b) 3y [n 2]
Figure 7: (c) 2 2y [2 + n]
Figure 8: HW 1.3 Decompose y [n] from Problem 2 into its even and odd parts ye [n] and yo [n], respectively. Plot these signals and show the symmetries of the plots that can be used to visually verify their parity. SOLUTION: Recall the formulas for even and odd ye [n] = 0.5(y [n] + y [n]) yo [n] = 0.5[y [n] y [n])
Figure 9: ye [n]
Figure 10: yo [n] HW 1.4 For each of the following discrete-time signals, if the signal is periodic then determine the fundamental period, the number of cycles per period, and the fundamental frequency. If the signal is not periodic, then show why.
n (a) x[n] = cos( 32 + 2 ) SOLUTION: n n x[n] = cos( 32 + 2 ) = sin( + 2 ) 2 Know sin(a + b) = sin(a) cos(b) + cos(a) sin(b) and that sin and cos are 2 periodic. There exist no n such that 2n is a multiple of 2 . Thus not periodic.
(b) x[n] = ej 7 n SOLUTION: Periodic if 4 n divided by 2 is rational. 7 4 2 n/2 = 7 , which is rational. 7 Thus the fundemental frequency is 27 . 5
3 (c) x[n] = cos( 4 n) + sin( 7 n) 5 SOLUTION: cos( 4 n) periodic every 5. 5 3 n) is periodic every 14. sin( 7 The LCM of 5 and 14 is 70, so the fundemental frequency is
35
(d) x[n] = ejn + ej 2n SOLUTION: Is not periodic because ej 2n is not periodic. 2 cannot be a multiple of 2. HW 1.5 For the following functions: (a) y [n] =
n k=n
x[a + k ], a Z
(b) y [n] = x(e|n| ) (c) y [n] = 0.5[x[n] + x[n 1]] (d) y [n] =
M 1 k=0
x[ n k ] , M N
Prove or disprove that they are: (i) memoryless (ii) invertible (iii) causal (iv) stable (v) time invariant (vi) linear (a) SOLUTION: (i) Quite clearly depends on previous input, so has memory. (ii) No, y [1] = x[a 1] + x[a] + x[a + 1] = y [1] and y [0] = x[a]. Generalize for all n, y [n] = y [n] = x[a n] + y [n 1] + x[a + n]. It is not possible to determine the x[a n] and x[a + n] factor. (iii) y [0] = x[a] Therefore not causal. (iv) Not stable. Suppose x[n] = 1 for all n. As n goes to innity, y [n] becomes unbounded.
nn0 (v) Shift y [n], y [n n0 ] = k =(nn0 ) x[a + k ] n Now shift x[n], k=n x[a + k n0 ] Both are not identical, thus not time-invariant. (vi) y [n] = n k=n (c0 x0 [a + k ] + c1 x1 [a + k ]) n = k=n c0 x0 [a + k ] + n k=n c1 x1 [a + k ] = c0 y0 [n] + c1 y1 [n] Thus the system is linear
(b) SOLUTION: (i) n = exp|n|| , n Z So not memoryless. (ii) No, exp|n| is always positive, so x[n] with n < 0 cannot be determined. (iii) y [0] = x[exp0 ] = x[1] Thus, not causal. (iv) Yes. If x[n] < B as n goes to innity, then y [n] < B as n goes to innity. (v) Shift y [n], y [n n0 ] = x[exp|nn0 | ] Now shift x[n], x[exp|n| n0 ] Not equal, thus not time-invariant. (vi) y [n] = (c0 x0 [exp|n| ] + c1 x1 [exp|n| ]) = c0 y0 [n] + c1 y1 [n] System is Linear. (c) SOLUTION: (i) Quite clearly depends on previous input, so has memory. (ii) Only if the initial condition of x[n] is given. y [n] = 0.5[x[n] + x[n 1]] Suppose there is an initial condition x0 at time M. y [M ] = 0.5[x[M ] + x[M 1]] = 0.5x[M ] At time M + 1, y [M + 1] = 0.5[x[M + 1] + x[M ]] = 0.5[x[M + 1] + y [M ]] x[M + 1] is solvable. (iii) Yes, causal. (iv) Yes, stable. If x[n] < B then y [n] < B by subsitition. (v) Shift y [n], y [n n0 ] = 0.5[x[n n0 ] + x[n n0 1]] Now shift x[n], 0.5[x[n n0 ] + x[n n0 1]] Both are identical, thus time-invariant. (vi) y [n] = c0 x0 [n] + c0 x0 x[n 1] + c1 x1 [n] + c1 x1 [n 1]) = c0 y0 [n] + c1 y1 [n] Thus the system is linear 7
(d) SOLUTION: (i) Quite clearly depends on previous input, so has memory. (ii) Only if the initial condition of x[n] is given. 1 M 1 y [n] y [n 1] = M k=0 x[n k ] k=0 x[n k 1] = x[n] x[n M ] So if x[n M ] is known, then it is invertible. (iii) Yes, causal. Only depends on past M. (iv) Yes, stable. If x[n] < B then y [n] < B by subsitition.
1 (v) Shift y [n], y [n n0 ] = M k=0 x[n k n0 ] 1 Now shift x[n], M k=0 x[n k n0 ] Both are identical, thus timeinvariant. 1 (vi) y [n] = M k=0 (c0 x0 [n k ] + c1 x1 [n k ]) M 1 M 1 k=0 c1 x1 [n k ] k=0 c0 x0 [n k ] + = c0 y0 [n] + c1 y1 [n] Thus the system is linear
