Books in black and white
 Main menu Share a book About us Home
 Books Biology Business Chemistry Computers Culture Economics Fiction Games Guide History Management Mathematical Medicine Mental Fitnes Physics Psychology Scince Sport Technics

# Detection, Estimation modulation theory part 1 - Vantress H.

Vantress H. Detection, Estimation modulation theory part 1 - Wiley & sons , 2001. - 710 p.
ISBN 0-471-09517-6 Previous << 1 .. 139 140 141 142 143 144 < 145 > 146 147 148 149 150 151 .. 254 >> Next Hi Zi2 ? Zo2. Ho
398 4.8 Problems
V2"fi(t) sin (uct + <t>i(t))
ł = 0,1
Fig. P4.7
Problem 4.4.9 (continuation) . Assume that the signal components on the two hypotheses are orthogonal.
1. Assuming that H0 is true, verify that
E(x'0) = Am + d2 cos â,
= d2 sin 0,
?(*l) = Ëň,
Ĺ(óä = 0,
where d2 é 2ET/Na and
Var (jci) = Var (>¦») = Var (*0 = Var (*) = d2.
2. Prove that
Zo2 + Am2 + d* 4- 2Amd2 cos â\
2d2
and
x |/o ^[(Am2 + d* + 2Amd2 cos g)^Z0I Ë^ę^ăő˛ßî. â) = § exp
3. Show that
Pr (*) =Pr 0|#0) = Pr(z0 < Ăő˛ßî) = j* MQM P*0h0AZo\Ho, 9)dZ0
x f Pzi\Hi,e(,Zi\Hi, 6) dZi9 Jz0
4. Prove that the inner two integrals can be rewritten as
Pr («|â) = Q(a, b) - ł exp /„(ai),
_ ^2Am
where
b =
[2(Am2 + d* + 2Amd2 cos 0)P
5. Check your result for the two special cases in which Am->0 and Am-^oo. Compare the resulting Pr («) for these two cases in the region where d is large.
Signals with Unwanted Parameters 399
Problem 4.4.10 (continuation). Error Probability, Binary Nonorthogonal Signals . When bandpass signals are not orthogonal, it is conventional to define their correlation in the following manner:
Ř A fi(t)e>*i«'
Aw ?/i(/y*e»>
which is a complex number.
1. Express the actual signals in terms of ft(t).
2. Express the actual correlation coefficient of two signals in terms of p.
3. Assume Am = 0 (this corresponds to a uniform density) and define the quantity
A = (1 - \p\*fK
Show that
Pr 0) = Vi - A, I Vi + a) - ł exp (-y)/o(^- |/>|).
Problem 4.4.11 (continuation). When pe(0) is nonuniform and the signals are nonorthogonal, the calculations are much more tedious. Set up the problem and then refer to  for the detailed manipulations.
Problem 4.4.12. M-ary PSK. Consider the M-ary PSK communication system in Problem 4.2.21. Assume that
- ŞŐĐ (Ë„ COS 0) ^ a _
Ěâ)- 2ňň /„(Am) ’ —
1. Find the optimum receiver.
2. Write an expression for the Pr (c).
Problem 4.4.13. ASK: Incoherent Channel . An ASK system transmits equally likely messages
Si(t) = Vő2Et f(t) cos coct, / = 1, 2,..., M, 0 < t < T,
where
ve, = a - ł)ä, fof20)dt = 1,
and
(M - 1)Ä Ä E.
The received signal under the ith hypothesis is
r(t) = V2Ei f{t) cos (o>c/ 4- â) + wit), 0 < t < T: Íł, ł = 1, 2,..., M,
where wit) is white noise iN0/2). The phase â is a random variable with a uniform density (0, 2tt).
1. Find the minimum Pr (e) receiver.
2. Draw the decision space and compute the Pr (e).
400 4.8 Problems
Problem 4.4.14. Asymptotic Behavior of Incoherent M-ary Systems . In the text we saw that the probability of error in a communication system using M orthogonal signals approached zero as M^oo as long as the rate in digits per second was less than Đ/No In 2 (the channel capacity) (see pp. 264-267). Use exactly the same model as in Example 4 on pp. 264-267. Assume, however, that the channel adds a random phase angle. Prove that exactly the same results hold for this case. (Comment. The derivation is somewhat involved. The result is due originally to Turin . A detailed derivation is given in Section 8.10 of .)
Problem 4.4.15 . Calculate the moment generating function, mean, and variance of the test statistic G = L2 + Ls2 for the random phase problem of Section 4.4.1 under the hypothesis Hlm
Problem 4.4.16 (continuation) . We can show that for d > 3 the equivalent test statistic
is approximately Gaussian (see Fig. 4.73a). Assuming that this is true, find the mean and variance of R.
Problem 4.4.17 (continuation) . Now use the result of Problem 4.4.16 to derive an approximate expression for the probability of detection. Express the result in terms of d and PF and show that PD is approximately a straight line when plotted versus d on probability paper. Compare a few points with Fig. 4.59. Evaluate the increase in d over the known signal case that is necessary to achieve the same performance.
Problem 4.4.18. Amplitude Estimation. We consider a simple estimation problem in which an unwanted parameter is present. The received signal is
r(t) = AVls(t,e) + w(t), where A is a nonrandom parameter we want to estimate (assume that it is nonnegative): s(t, 0) = f(t) cos [<*>ct + ô(0 + 0], Previous << 1 .. 139 140 141 142 143 144 < 145 > 146 147 148 149 150 151 .. 254 >> Next 