Difference between revisions of "Aufgaben:Exercise 3.9Z: Convolution of Gaussian Pulses"

From LNTwww
 
Line 3: Line 3:
 
}}
 
}}
  
[[File:P_ID544__Sig_Z_3_9.png|right|frame|Gaussian pulses   $x(t)$  and  $h(t)$]]
+
[[File:P_ID544__Sig_Z_3_9.png|right|frame|Gaussian pulses   $x(t)$,   $h(t)$]]
 
The convolution result of two Gaussian functions is to be determined.  We consider  
 
The convolution result of two Gaussian functions is to be determined.  We consider  
 
*a Gaussian input pulse  ${x(t)}$  with amplitude $x_0 = 1\,\text{V}$ and  "equivalent pulse duration"  $\Delta t_x = 4 \,\text{ms}$,  as well as  
 
*a Gaussian input pulse  ${x(t)}$  with amplitude $x_0 = 1\,\text{V}$ and  "equivalent pulse duration"  $\Delta t_x = 4 \,\text{ms}$,  as well as  
Line 50: Line 50:
 
:$$X( f ) = x_0  \cdot \Delta t_x  \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x  \hspace{0.05cm}\cdot \hspace{0.05cm} f} \right)^2 } , \hspace{0.5cm}H(f) = {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_h  \hspace{0.05cm}\cdot \hspace{0.05cm}f} \right)^2 } .$$
 
:$$X( f ) = x_0  \cdot \Delta t_x  \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x  \hspace{0.05cm}\cdot \hspace{0.05cm} f} \right)^2 } , \hspace{0.5cm}H(f) = {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_h  \hspace{0.05cm}\cdot \hspace{0.05cm}f} \right)^2 } .$$
 
*The values we are looking for are  
 
*The values we are looking for are  
:$$X(f = 0)\;\underline{ = 4 \,\text{mV/Hz}},$$
+
:$$X(f = 0)\;\underline{ = 4 \,\text{mV/Hz}}, \hspace{0.5cm}
:$$H(f = 0)\; \underline{= 1}.$$
+
H(f = 0)\; \underline{= 1}.$$
  
  
[[File:P_ID589__Sig_Z_3_9_b_neu.png|right|frame|Convolution result for „$\rm Gauss  \ast  Gauss$”]]
+
[[File:P_ID589__Sig_Z_3_9_b_neu.png|right|frame|Gaussian spektra  $X(f)$,     $Y(f)$     –     Gaussian pulses  $x(t)$,     $y(t)$]]
'''(2)'''   Convolution in the time domain corresponds to multiplication in the frequency domain:
+
'''(2)'''   Convolution in time domain corresponds to multiplication in frequency domain:
 
:$$Y(f) = X(f) \cdot H(f) = x_0  \cdot \Delta t_x  \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x^2  + \Delta t_h^2 } \right)f^2 } .$$
 
:$$Y(f) = X(f) \cdot H(f) = x_0  \cdot \Delta t_x  \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x^2  + \Delta t_h^2 } \right)f^2 } .$$
 
*With the abbreviation  $\Delta t_y = (\Delta t_x^2 + \Delta t_h^2)^{1/2} = 5\, \text{ms}$  one can write for this:
 
*With the abbreviation  $\Delta t_y = (\Delta t_x^2 + \Delta t_h^2)^{1/2} = 5\, \text{ms}$  one can write for this:
Line 69: Line 69:
 
*This gives:
 
*This gives:
 
:$$y(t) = x(t) * h(t) = x_0  \cdot \frac{\Delta t_x }{\Delta t_y } \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {t/\Delta t_y } \right)^2 } .$$
 
:$$y(t) = x(t) * h(t) = x_0  \cdot \frac{\Delta t_x }{\Delta t_y } \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {t/\Delta t_y } \right)^2 } .$$
*The maximum value of the signal  ${y(t)}$  is also at  $t = 0$  and is  $y_0 \hspace{0.15cm}\underline{= 0.8 \text{ V} }$.  
+
*The maximum value of the signal  ${y(t)}$  is also at   $t = 0$  and is   $y_0 \hspace{0.15cm}\underline{= 0.8 \text{ V} }$.  
*The equivalent pulse duration results in  $\Delta t_y \hspace{0.15cm}\underline{= 5 \text{ ms}}$  (see above picture, right sketch).  
+
*The equivalent pulse duration results in  $\Delta t_y \hspace{0.15cm}\underline{= 5 \text{ ms}}$  (see above graphic, right sketch).  
 
*This means:  The Gaussian  ${H(f)}$  causes the output pulse  ${y(t)}$  to be smaller and wider than the input pulse  ${x(t)}$ .
 
*This means:  The Gaussian  ${H(f)}$  causes the output pulse  ${y(t)}$  to be smaller and wider than the input pulse  ${x(t)}$ .
*The pulse shape remains Gaussian because:   '''Gaussian convoluted with Gaussian always results in Gaussian!'''
+
*The pulse shape remains Gaussian.   Because:   '''Gaussian convoluted with Gaussian always results in Gaussian!'''
 
{{ML-Fuß}}
 
{{ML-Fuß}}
  

Latest revision as of 16:45, 29 April 2021

Gaussian pulses  $x(t)$,   $h(t)$

The convolution result of two Gaussian functions is to be determined.  We consider

  • a Gaussian input pulse  ${x(t)}$  with amplitude $x_0 = 1\,\text{V}$ and  "equivalent pulse duration"  $\Delta t_x = 4 \,\text{ms}$,  as well as
  • a likewise Gaussian impulse response  ${h(t)}$, which has the  "equivalent pulse duration"  $\Delta t_h = 3 \,\text{ms}$ :
$$x( t ) = x_0 \cdot {\rm{e}}^{ - {\rm{\pi }}( {t/\Delta t_x } )^2 } ,$$
$$h( t ) = \frac{1}{\Delta t_h } \cdot {\rm{e}}^{ - {\rm{\pi }}( {t/\Delta t_h } )^2 } .$$

The output signal  ${y(t)} = {x(t)} ∗{h(t)}$  is sought, whereby the diversions via the spectral functions is to be taken.




Hint:



Questions

1

Give the spectral functions  ${X(f)}$  and  ${H(f)}$  an.  Which values result for  $f = 0$?

$X(f = 0)\ = \ $

 $\text{mV/Hz}$
$H(f = 0)\ = \ $

2

Calculate the spectral function  ${Y(f)}$  of the output signal.  What is the spectral value at  $f = 0$?

$Y(f = 0)\ = \ $

 $\text{mV/Hz}$

3

Calculate the output pulse  ${y(t)}$.  What values result for the amplitude  $y_0 = y(t = 0)$  and the equivalent pulse duration  $\Delta t_y$?

$y_0\ = \ $

 $\text{V}$
$\Delta t_y\ = \ $

 $\text{ms}$


Solution

(1)  By Fourier transformation one obtains:

$$X( f ) = x_0 \cdot \Delta t_x \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x \hspace{0.05cm}\cdot \hspace{0.05cm} f} \right)^2 } , \hspace{0.5cm}H(f) = {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_h \hspace{0.05cm}\cdot \hspace{0.05cm}f} \right)^2 } .$$
  • The values we are looking for are
$$X(f = 0)\;\underline{ = 4 \,\text{mV/Hz}}, \hspace{0.5cm} H(f = 0)\; \underline{= 1}.$$


Gaussian spektra  $X(f)$,     $Y(f)$     –     Gaussian pulses  $x(t)$,     $y(t)$

(2)  Convolution in time domain corresponds to multiplication in frequency domain:

$$Y(f) = X(f) \cdot H(f) = x_0 \cdot \Delta t_x \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_x^2 + \Delta t_h^2 } \right)f^2 } .$$
  • With the abbreviation  $\Delta t_y = (\Delta t_x^2 + \Delta t_h^2)^{1/2} = 5\, \text{ms}$  one can write for this:
$$Y(f) = x_0 \cdot \Delta t_x \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_y \hspace{0.05cm}\cdot \hspace{0.05cm} f} \right)^2 } .$$
  • At frequency  $f = 0$ , the spectral values at the input and output of the Gaussian filter are equal, so:
$$Y(f = 0) \;\underline{= 4 \text{ mV/Hz}}.$$
  • The function curve of  ${Y(f)}$  is narrower than  ${X(f)}$  and narrower than  ${H(f)}$.


(3)  The following Fourier correspondence holds:

$${\rm{e}}^{ - {\rm{\pi }}\left( {\Delta t_y \hspace{0.05cm}\cdot \hspace{0.05cm} f} \right)^2 }\bullet\!\!\!-\!\!\!-\!\!\!-\!\!\circ\, \frac{1}{\Delta t_y } \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {t/\Delta t_y } \right)^2 } .$$
  • This gives:
$$y(t) = x(t) * h(t) = x_0 \cdot \frac{\Delta t_x }{\Delta t_y } \cdot {\rm{e}}^{ - {\rm{\pi }}\left( {t/\Delta t_y } \right)^2 } .$$
  • The maximum value of the signal  ${y(t)}$  is also at   $t = 0$  and is   $y_0 \hspace{0.15cm}\underline{= 0.8 \text{ V} }$.
  • The equivalent pulse duration results in  $\Delta t_y \hspace{0.15cm}\underline{= 5 \text{ ms}}$  (see above graphic, right sketch).
  • This means:  The Gaussian  ${H(f)}$  causes the output pulse  ${y(t)}$  to be smaller and wider than the input pulse  ${x(t)}$ .
  • The pulse shape remains Gaussian.   Because:   Gaussian convoluted with Gaussian always results in Gaussian!