Scribd
Upload
90 views

Detection and Estimation HW

This document contains a practice exam for a detection and estimation theory midterm. It includes 4 multiple choice questions about hypothesis testing, detection problems, and optimal detectors. Specifically, it asks about hypothesis testing to detect groundwater pollution, detecting defective RAM chips from different batches, determining which communication channel is being used based on signal and noise models, and true/false questions about detection theory concepts.

Uploaded by

demet
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
90 views

Detection and Estimation HW

This document contains a practice exam for a detection and estimation theory midterm. It includes 4 multiple choice questions about hypothesis testing, detection problems, and optimal detectors. Specifically, it asks about hypothesis testing to detect groundwater pollution, detecting defective RAM chips from different batches, determining which communication channel is being used based on signal and noise models, and true/false questions about detection theory concepts.

Uploaded by

demet
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

University of Illinois at Chicago

Department of Electrical and Computer Engineering

EC 531: Detection and Estimation Theory


Spring 2009
Practice for Midterm 1

1. (30 points) You are a lawyer. In a community near Chicago, there is the suspicion that the ground-water has
been polluted by a large chemical company. Two hypotheses exist: under hypothesis H1 the ground- water is
polluted and will cause cancer, and under hypothesis H0 the ground-water is fine and drinkable. Assume that
the results of tests on the ground-water obtained from a series of pilot wells in the area are summarized by the
value of a random variable X . If the ground-water is polluted the random variable X has the following pdf
p(x; H1 ) = xe−x u(x) where u(x) is the unit step function (0 for x < 0 and 1 for x ≥ 0). In the case that the
ground-water is fine the random variable X obeys the following pdf: p(x; H0 ) = e−x u(x). You must decide, based
on the value of X that is obtained from the well tests, whether or not to initiate legal action against the chemical
company.

(a) Design a hypothesis testing rule (find decision regions) such that, PD , the probability that you decide to sue
given that the ground-water is polluted is present is maximized subject to the constraint that the probability of
suing given the ground-water is fine is 0.1.

(b) Find the resulting value of PD .

(c) Assuming the probability that the water is polluted is 0.1, find the test that will yield the minimum probability
of error.

(d) What is the resulting probability of error for part (c)?

1
2. (20 points) When a RAM chip is selected from a particular production batch there is a probability α that the
RAM chip is defective. Thus, k, the number of defective RAM chips out of N independently selected ones is a
binomial random variable:
N !αK (1 − α)N −K
P r[k = K] =
K!(N − K)!
Suppose that we receive a shipment of N RAM chips which either all came from batch 0 (probability of defect
α = α0 ) or all came from batch 1 (probability of defect α = α1 ), with a priori probabilities P0 and P1 , respectively.
Assume 0 < α0 < α1 < 1. We wish to determine which production batch our RAM chips came from by testing
how many of the N units are defective. Let x denote the number of defective RAM chips out of the total of N
RAM chips.

(a) Find the minimum probability of error rule for deciding whether the received RAM chips were from batch 0
or from batch 1.

(b) Suppose N = 2 and plot the ROC. (HINT: since x takes on discrete values, the ROC consists of a number of
points).

2
3. (30 points) Suppose we want to determine which of two channels is being used for communications. We observe
the output of the channel x(n) for n = 0, 1, · · · , N − 1. We know that

If channel 1 is used: x(n) = a1 s1 (n) + w1 (n)


If channel 2 is used: x(n) = a2 s2 (n) + w2 (n)

(a) Suppose a1 , a2 , s1 (n), s2 (n) are all known and that w1 ∼ N (0, σ 2 I), w2 ∼ N (0, σ 2 I). Describe the detector
that minimizes the probability of error.

(b) Suppose a1 , a2 , s1 (n), s2 (n) are all known and that w1 ∼ N (0, CW1 ), w2 ∼ N (0, CW2 ). Describe the detector
that minimizes the probability of error (you do not need to simplify).

(c) Suppose a1 , a2 are unknown, s1 (n), s2 (n) are known and w1 ∼ N (0, CW! ), w2 ∼ N (0, CW2 ). Determine the
generalized likelihood ratio test (again, you do not need to simplify).

3
4. (20 points) TRUE or FALSE:

(a) The Neyman-Pearson test yields the best probability of detection for a given probability of false alarm regardless
of the noise statistics.

(b) When x ∼ N (0, σi2 ) under Hi , (i = 0, 1), the decision region depend only the ratio of σ02 /σ12 .

(c) The Bayesian risk formulation captures the maximum a posteriori, maximum likelihood and Neyman-Pearon
detectors.

(d) When detecting which of multiple known signals was sent in white Gaussian noise the test statistic always
boils down to a minimum distance test.

(e) The performance of a matched filter in colored noise does not depend on the signal shape.

(f) When detecting a random Gaussian signal in Gaussian noise, the test statistic is linear in the received data.

(g) Detecting a sinusoid of unknown amplitude, phase and frequency in white Gaussian noise may be cast in the
linear model.

(h) The generalized likelihood ratio test outperforms the Rao and Wald tests when detecting a signal with unknown
parameters when you have a finite number of samples.

You might also like