Matched Filter

1- Consider the signal S(t) shown in Figure 1.

i- Determine the impulse response of a filter matched to this signal and sketch it as a
function of time.
ii- Plot the matched filter output as a function of time.
iii- What is the peak value of the output?

2- Find the output of the matched filter and determine the maximum
value of output signal to noise ratio (S/N) if the input signal S(t) is
rectangular pulse of amplitude A and duration T.

3- Let S(t) is a signal passed through an Additive White Gaussian Noise (AWGN) channel
𝑁𝑜
with noise power spectral density PSD = . if we use a matched filter receiver with
2
impulse response h(t).
i. Drive an expression for optimum impulse response hopt(t) of a matched filter
receiver that is used as linear filter for maximizing the output signal- to-noise ratio
in the presence of Additive White Gaussian Noise.
𝑡 −3
ii. Consider the signal s(t) = 2 𝑟𝑒𝑐𝑡 ( ) Plot the matched filter output as a
6
function of time.
4- Let s(t) is a signal passed through an Additive White Gaussian Noise AWGN channel with
𝑁𝑜
noise power spectral density PSD = . if we use a matched filter receiver with optimum
2
impulse response.
i. Drive an expression for maximize output signal- to-noise ratio 𝜂𝑚𝑎𝑥
𝑡 −4
ii. Let s(t) = 𝑟𝑒𝑐𝑡 ( ) and 𝑁𝑜 = 10−5 𝑤𝑎𝑡𝑡/𝐻𝑧.
8
Calculate Maximum output Signal- to-Noise Ratio 𝜂𝑚𝑎𝑥 .
5- If S(t): transmitted input signal with period T and w(t): Additive White Gaussian Noise are
input to a matched filter receiver with optimum impulse response 𝐻𝑜𝑝𝑡 (𝑓) =
𝑘 𝑆 ∗ (𝑓) 𝑒 −𝑗2𝜋𝑓𝑇 .
𝑇
i. Show that the matched filter output signal 𝑦(𝑇) = 𝑘𝐸𝑏 + 𝑘 ∫0 𝑤(𝜏) 𝑆(𝜏) 𝑑𝜏.
Where 𝐸𝑏 is the bit energy of a binary digital signal
 ii. If 𝐸𝑏 = 10−5 𝑤𝑠 and the power spectral density (PSD) of additive white noise
(AWGN) 𝑁𝑜 = 10−6 𝑤𝑎𝑡𝑡/𝐻𝑧 Calculate in dB Maximum output Signal- to-
Noise Ratio 𝜂𝑚𝑎𝑥 .
6- Drive the optimum matched filter frequency response 𝐻𝑜 (𝑤) that maximize the output signal to
nose ratio SNR when the noise not Whit Gawssian Noise.
Advanced Problems
7-

8-

Schwarz's inequality
𝑖𝑓𝑓 𝜓1 (𝑥) = 𝑘 ′ 𝜓2 ∗ (𝑥) ; 𝑘 ′ is a constant
1- Problem 1 sol
2- Problem 2 sol
 3- Problem 3 sol

𝑦(𝑡)
𝑦(𝑡) = 𝑜 (𝑡)∗+
= 𝑆𝑠(𝑡) 𝑛(𝑡)+ 𝑤(𝑡)
ℎ(𝑡)
∗ ℎ(𝑡)

i.

|𝑆𝑜 (𝑡)|2
𝜂=
𝐸[𝑛2 (𝑡)]
∞
𝑆𝑜 (𝑡) = 𝐼𝐹𝑇 {𝑠(𝑓)𝐻(𝑓)} 𝐸[𝑛2 (𝑡)] = ∫−∞ 𝑆𝑛 (𝑓) 𝑑𝑓
∞ 2
|∫−∞ 𝑠(𝑓)𝐻(𝑓)𝑒 𝑗2𝜋𝑓𝑇 𝑑𝑓 |
𝜂= ∞ 𝑁
∫−∞ 2𝑜 |𝐻(𝑓)|2 𝑑𝑓

By using Schwarz's inequality :

𝑁 1
𝜓1 (𝑥) = √ 2𝑜 𝐻(𝑓) 𝜓2 (𝑥) = 𝑆(𝑓) 𝑒 𝑗2𝜋𝑓𝑇
𝑁
√ 𝑜
2

𝑁𝑜 𝑆 ∗ (𝑓) 𝑒 −𝑗2𝜋𝑓𝑇
√ 𝐻𝑜𝑝𝑡 (𝑓) = 𝑘 ′ 𝐻𝑜𝑝𝑡 (𝑓) = 𝑘 𝑆 ∗ (𝑓) 𝑒 −𝑗2𝜋𝑓𝑇
2 𝑁
√ 𝑜
2

𝐻𝑜𝑝𝑡 (𝑓) = 𝑘 𝑆(−𝑓) 𝑒 −𝑗2𝜋𝑓𝑇 𝐻𝑜𝑝𝑡 (𝑓) = 𝑘 𝑆(−𝑓) 𝑒 −𝑗2𝜋𝑓𝑇

𝑡 −3
ii. For s(t) = 2 𝑟𝑒𝑐𝑡 ( ) , T = 6 and A= 2
6
 4- Problem 4 sol

𝑦(𝑡) = 𝑠(𝑡) ∗ ℎ(𝑡) + 𝑤(𝑡)

∗ ℎ(𝑡)
𝑦(𝑡) = 𝑆𝑜 (𝑡) + 𝑛(𝑡)

i.

|𝑆𝑜 (𝑡)|2
𝜂=
𝐸[𝑛2 (𝑡)]
∞
𝑆𝑜 (𝑡) = 𝐼𝐹𝑇 {𝑠(𝑓)𝐻(𝑓)} 𝐸[𝑛2 (𝑡)] = ∫−∞ 𝑆𝑛 (𝑓) 𝑑𝑓
∞ 2
|∫−∞ 𝑠(𝑓)𝐻(𝑓)𝑒 𝑗2𝜋𝑓𝑇 𝑑𝑓 |
𝜂= ∞ 𝑁
∫−∞ 2𝑜 |𝐻(𝑓)|2 𝑑𝑓

By using Schwarz's inequality:

𝑁𝑜 1
𝜓1 (𝑥) = √ 𝐻(𝑓) 𝜓2 (𝑥) = 𝑆(𝑓) 𝑒 𝑗2𝜋𝑓𝑇
2 𝑁
√ 𝑜
2

∞ 2 𝑁𝑜 ∞
∫−∞ 𝑁 |𝑠(𝑓)|2 𝑑𝑓 2 ∫−∞
|𝐻(𝑓)|2 𝑑𝑓
𝑜
𝜂≤
𝑁𝑜 ∞
2 ∫−∞|𝐻(𝑓)|2 𝑑𝑓

2 ∞ 2 𝐸𝑔
𝜂𝑚𝑎𝑥 = 𝑁 ∫−∞ |𝑠(𝑓)|2 𝑑𝑓 𝜂𝑚𝑎𝑥 =
𝑜 𝑁𝑜
𝑡 −4
ii. For s(t) = 𝑟𝑒𝑐𝑡 ( ) and 𝑁𝑜 = 10−5 𝑤𝑎𝑡𝑡/𝐻𝑧.
8
 8
𝐸𝑔 = ∫ |𝑠(𝑡)|2 𝑑𝑡 = 12 𝑥 8 = 8 𝑤𝑠
0

2 𝐸𝑔 2 𝑥 8
𝜂𝑚𝑎𝑥 = = −5 = 16𝑥105
𝑁𝑜 10

𝜂max 𝑑𝐵 = 10 log (16𝑥105 ) = 62.04 dB

5- Problem 5 sol
i.

ii. .
2 𝐸𝑔 2 𝑥 10−5
𝜂𝑚𝑎𝑥 = = = 20
𝑁𝑜 10−5
𝜂max 𝑑𝐵 = 10 log (20) = 13 dB
6- Problem 6 sol