Difference between revisions of "Aufgaben:Exercise 3.5Z: Antenna Areas"
m (Text replacement - "rms value" to "standard deviation") |
|||
(12 intermediate revisions by 3 users not shown) | |||
Line 1: | Line 1: | ||
− | {{quiz-Header|Buchseite= | + | {{quiz-Header|Buchseite=Theory_of_Stochastic_Signals/Uniformly_Distributed_Random_Variables |
}} | }} | ||
− | [[File:P_ID188__Sto_Z_3_5.png|right|frame| | + | [[File:P_ID188__Sto_Z_3_5.png|right|frame|Two antenna areas: $K$ and $G$]] |
− | + | We first consider – as sketched in the image above – a receiving antenna serving a circular area $K$. It is assumed that this antenna can detect all signals incident at different angles $\alpha$ equally well: | |
− | * | + | *According to the sketch, the angle $\alpha$ refers to the $x$–axis. |
− | * | + | *The value $\alpha = 0$ therefore means that the signal is moving towards the antenna in the direction of the negative $x$–axis. |
− | + | Further we assume: | |
− | * | + | *The range of values of the angle of incidence $\alpha$ with this definition $-\pi < \alpha \le +\pi$. |
− | * | + | *There are very many users in the coverage area whose positions $(x, y)$ are "statistically distributed" over the area $K$. |
− | + | From subtask '''(5)''' we assume the coverage area $G$ outlined below. | |
− | * | + | *Because of an obstacle, the $x$–coordinate of all participants must now be greaterö&space;than $-R/2$. |
− | * | + | *Also in the coverage area $G$ the subscribers would again be "statistically distributed". |
Line 22: | Line 22: | ||
− | + | Hint: | |
− | + | *The exercise belongs to the chapter [[Theory_of_Stochastic_Signals/Uniformly_Distributed_Random_Variables|Uniformly Distributed Random Variables]]. | |
− | |||
− | |||
− | * | ||
− | === | + | ===Questions=== |
<quiz display=simple> | <quiz display=simple> | ||
− | { | + | {What is the PDF $f_\alpha(\alpha)$ for the area $K$? What PDF–value results for $\alpha = 0$? |
|type="{}"} | |type="{}"} | ||
$f_\alpha(\alpha = 0) \ = \ $ { 0.159 3% } | $f_\alpha(\alpha = 0) \ = \ $ { 0.159 3% } | ||
− | { | + | {Which of the two statements is correct? Note in particular also the asymmetric definition range of $-\pi < \alpha \le +\pi$. |
|type="()"} | |type="()"} | ||
− | + | + | + The expected value is ${\rm E}[\alpha] = 0$. |
− | - | + | - The expected value is ${\rm E}[\alpha] \ne 0$. |
− | { | + | {What value results for the standard deviation of the random variable $\alpha$ in the area $K$? |
|type="{}"} | |type="{}"} | ||
− | $\sigma_\alpha \ = | + | $\sigma_\alpha \ = \ $ { 1.814 3% } |
− | { | + | {What is the probability that in area $K$ the antenna locates a user at an angle between $\pm45^\circ$ ? |
|type="{}"} | |type="{}"} | ||
− | ${\rm Pr}( | + | ${\rm Pr}(-π/4 ≤ α ≤ +π/4) \ = \ $ { 25 3% } $\ \%$ |
− | { | + | {Now let's consider the coverage area $G$. In which area $-\alpha_0 \le \alpha \le +\alpha_0$ does the PDF $f_\alpha(\alpha)$ have a constant value? |
|type="{}"} | |type="{}"} | ||
− | $\alpha_0 \ = \ $ | + | $\alpha_0 \ = \ $ { 2.094 3% } $ \ \rm rad$ |
− | $\alpha_0 \ = \ $ | + | $\alpha_0 \ = \ $ { 120 3% } $ \ \rm degrees$ |
− | { | + | {What statements are now valid with respect to $f_\alpha(\alpha)$ in the range $|\alpha| > \alpha_0$ ? |
|type="[]"} | |type="[]"} | ||
− | - | + | - The PDF has the same course "outside" as "inside". |
− | - | + | - The PDF is "outside" identically zero. |
− | + | + | + The PDF decreases towards the edges in this area. |
− | - | + | - The PDF increases towards the edges in this area. |
− | { | + | {Calculate for the area $G$ the probability that the antenna locates a user at an angle between $\pm 45^\circ$ . |
|type="{}"} | |type="{}"} | ||
− | ${\rm Pr}( | + | ${\rm Pr}(-π/4 ≤ α ≤ +π/4) \ = \ $ { 31.1 3% } $\ \%$ |
− | { | + | {What is now the PDF value at the position $\alpha = 0$? |
|type="{}"} | |type="{}"} | ||
− | $f_\alpha(\alpha = | + | $f_\alpha(\alpha = 0) \ = \ $ { 0.198 3% } |
</quiz> | </quiz> | ||
− | === | + | ===Solution=== |
{{ML-Kopf}} | {{ML-Kopf}} | ||
− | '''(1)''' | + | '''(1)''' There is a uniform distribution and it is true for the PDF in the range $-\pi < \alpha \le +\pi$: |
+ | :$$f_\alpha(\alpha)={\rm 1}/({\rm 2\cdot \pi}).$$ | ||
+ | * For $\alpha = 0$ this gives – as for all allowed values also – the PDF value :$$f_\alpha(\alpha =0) \hspace{0.15cm}\underline{=0.159}.$$ | ||
+ | |||
+ | |||
+ | '''(2)''' It holds ${\rm E}\big[\alpha\big] = 0$ ⇒ <u>Answer 1</u>. | ||
+ | *It has no effect that $\alpha = +\pi$ is allowed, but $\alpha = -\pi$ is excluded. | ||
− | |||
− | '''(3)''' | + | '''(3)''' For the variance of the angle of incidence $\alpha$ holds: |
− | :$$\sigma_{\alpha}^{\rm 2}=\int_{-\rm\pi}^{\rm\pi}\hspace{-0.1cm}\it\alpha^{\rm 2}\cdot \it f_{\alpha}(\alpha)\,\,{\rm d} \alpha=\frac{\rm 1}{\rm 2\cdot\it \pi}\cdot \frac{\alpha^{\rm 3}}{\rm 3}\Bigg|_{\rm -\pi}^{\rm\pi}=\frac{\rm 2\cdot\pi^{3}}{\rm 2\cdot\rm \pi\cdot \rm 3}=\frac{\rm \pi^2}{\rm 3} = \rm 3.29. \hspace{0.5cm}\Rightarrow \hspace{0.5cm}\sigma_{\alpha}\hspace{0.15cm}\underline{=1.814}.$$ | + | :$$\sigma_{\alpha}^{\rm 2}=\int_{-\rm\pi}^{\rm\pi}\hspace{-0.1cm}\it\alpha^{\rm 2}\cdot \it f_{\alpha}(\alpha)\,\,{\rm d} \alpha=\frac{\rm 1}{\rm 2\cdot\it \pi}\cdot \frac{\alpha^{\rm 3}}{\rm 3}\Bigg|_{\rm -\pi}^{\rm\pi}=\frac{\rm 2\cdot\pi^{3}}{\rm 2\cdot\rm \pi\cdot \rm 3}=\frac{\rm \pi^2}{\rm 3} = \rm 3. 29. \hspace{0.5cm}\Rightarrow \hspace{0.5cm}\sigma_{\alpha}\hspace{0.15cm}\underline{=1.814}.$$ |
− | '''(4)''' | + | '''(4)''' Since the given section of the circle is exactly one quarter of the total circle area, the probability we are looking for is |
− | :$${\rm Pr}( | + | :$${\rm Pr}(-π/4 ≤ α ≤ +π/4)\hspace{0.15cm}\underline{=25\%}.$$ |
− | [[File: | + | [[File:EN_Sto_Z_3_5_e.png|right|frame|The area $G$]] |
− | '''(5)''' | + | '''(5)''' From simple geometrical ¨considerations (right-angled triangle, marked dark blue in the adjacent sketch) one obtains the equation of determination for the angle $\alpha_0$: |
:$$\cos(\pi-\alpha_{\rm 0}) = \frac{R/ 2}{R}={\rm 1}/{\rm 2}\hspace{0.5cm}\Rightarrow\hspace{0.5cm}\rm\pi-\it\alpha_{\rm 0}=\frac{\rm\pi}{\rm 3} \hspace{0.2cm}\rm( 60^{\circ}).$$ | :$$\cos(\pi-\alpha_{\rm 0}) = \frac{R/ 2}{R}={\rm 1}/{\rm 2}\hspace{0.5cm}\Rightarrow\hspace{0.5cm}\rm\pi-\it\alpha_{\rm 0}=\frac{\rm\pi}{\rm 3} \hspace{0.2cm}\rm( 60^{\circ}).$$ | ||
− | + | *It follows $\alpha_0 = \pi/3\hspace{0.15cm}\underline{=2.094}.$ | |
− | + | *This corresponds $\alpha_0 \hspace{0.15cm}\underline{=120^\circ}$. | |
− | '''(6)''' | + | |
− | * | + | |
− | * | + | |
− | * | + | '''(6)''' Correct is <u>the suggested solution 3</u>: |
− | * | + | *The PDF $f_\alpha(\alpha)$ is f for a given angle $\alpha$ directly proportional to the distance $A$ between the antenna and the boundary line. |
+ | *For $\alpha = \pm 2\pi/3 = \pm 120^\circ$ against $A = R$, for $\alpha \pm \pi = \pm 180^\circ$ against $A = R/2$. | ||
+ | *In between the distance becomes successively smaller. This means: The PDF decreases towards the boundary. | ||
+ | *The decrease follows the following course: | ||
:$$\it A=\frac{\it R/\rm 2}{\rm cos(\rm \pi-\it\alpha)}.$$ | :$$\it A=\frac{\it R/\rm 2}{\rm cos(\rm \pi-\it\alpha)}.$$ | ||
− | |||
− | |||
− | + | '''(7)''' The area $G$ can be calculated from the sum of the $240^\circ$–sector and the triangle formed by the vertices $\rm UVW$ : | |
− | :$$\rm Pr(\rm -\pi/4\le\it\alpha\le+\rm\pi/4)=\frac{\it F}{\it G}=\frac{1/4}{2/3+{\rm sin(60^{\circ})}/({\rm 2\pi})}=\frac{\rm 0.25}{\rm 0.805}\hspace{0.15cm}\underline{=\rm 0. | + | :$$G=\frac{\rm 2}{\rm 3}\cdot \it R^{\rm 2}\cdot{\rm \pi} \ {\rm +} \ \frac{\it R}{\rm 2}\cdot \it R\cdot \rm sin(\rm 60^{\circ}) = \it R^{\rm 2}\cdot \rm\pi\cdot (\frac{\rm 2}{\rm 3}+\frac{\rm \sqrt{3}}{\rm 4\cdot\pi}).$$ |
+ | |||
+ | *The probability we are looking for is given by the ratio of the areas $F$ and $G$ (see sketch): | ||
+ | :$$\rm Pr(\rm -\pi/4\le\it\alpha\le+\rm\pi/4)=\frac{\it F}{\it G}=\frac{1/4}{2/3+{\rm sin(60^{\circ})}/({\rm 2\pi})}=\frac{\rm 0.25}{\rm 0.805}\hspace{0.15cm}\underline{=\rm 31.1\%}.$$ | ||
+ | |||
+ | *Although nothing has changed from point '''(4)''' at the area $F$ the probability now becomes larger by a factor $1/0.805 ≈ 1.242$ due to the smaller area $G$ . | ||
− | |||
− | '''(8)''' | + | '''(8)''' Since the overall PDF area is constantly equal $1$ and the PDF decreases at the boundaries, it must have a larger value in the range $|\alpha| < 2\pi/3$ than in '''(1)'''. |
+ | * With the results from '''(1)''' and '''(7)''' holds: | ||
:$$f_{\alpha}(\alpha = 0)=\frac{1/(2\pi)}{2/3+{\rm sin(\rm 60^{\circ})}/({\rm 2\pi})} = \frac{\rm 1}{{\rm 4\cdot\pi}/{\rm 3}+\rm sin(60^{\circ})}\hspace{0.15cm}\underline{\approx \rm 0.198}.$$ | :$$f_{\alpha}(\alpha = 0)=\frac{1/(2\pi)}{2/3+{\rm sin(\rm 60^{\circ})}/({\rm 2\pi})} = \frac{\rm 1}{{\rm 4\cdot\pi}/{\rm 3}+\rm sin(60^{\circ})}\hspace{0.15cm}\underline{\approx \rm 0.198}.$$ | ||
− | + | *Like the probability in '''(7)''' also simultaneously the PDF value in the range $|\alpha| < 2\pi/3$ increases by a factor $1.242$ as the coverage area becomes smaller. | |
{{ML-Fuß}} | {{ML-Fuß}} | ||
Line 127: | Line 135: | ||
− | [[Category: | + | [[Category:Theory of Stochastic Signals: Exercises|^3.4 Uniformly Distributed Random Variable^]] |
Latest revision as of 12:11, 17 February 2022
We first consider – as sketched in the image above – a receiving antenna serving a circular area $K$. It is assumed that this antenna can detect all signals incident at different angles $\alpha$ equally well:
- According to the sketch, the angle $\alpha$ refers to the $x$–axis.
- The value $\alpha = 0$ therefore means that the signal is moving towards the antenna in the direction of the negative $x$–axis.
Further we assume:
- The range of values of the angle of incidence $\alpha$ with this definition $-\pi < \alpha \le +\pi$.
- There are very many users in the coverage area whose positions $(x, y)$ are "statistically distributed" over the area $K$.
From subtask (5) we assume the coverage area $G$ outlined below.
- Because of an obstacle, the $x$–coordinate of all participants must now be greaterö&space;than $-R/2$.
- Also in the coverage area $G$ the subscribers would again be "statistically distributed".
Hint:
- The exercise belongs to the chapter Uniformly Distributed Random Variables.
Questions
Solution
- $$f_\alpha(\alpha)={\rm 1}/({\rm 2\cdot \pi}).$$
- For $\alpha = 0$ this gives – as for all allowed values also – the PDF value :$$f_\alpha(\alpha =0) \hspace{0.15cm}\underline{=0.159}.$$
(2) It holds ${\rm E}\big[\alpha\big] = 0$ ⇒ Answer 1.
- It has no effect that $\alpha = +\pi$ is allowed, but $\alpha = -\pi$ is excluded.
(3) For the variance of the angle of incidence $\alpha$ holds:
- $$\sigma_{\alpha}^{\rm 2}=\int_{-\rm\pi}^{\rm\pi}\hspace{-0.1cm}\it\alpha^{\rm 2}\cdot \it f_{\alpha}(\alpha)\,\,{\rm d} \alpha=\frac{\rm 1}{\rm 2\cdot\it \pi}\cdot \frac{\alpha^{\rm 3}}{\rm 3}\Bigg|_{\rm -\pi}^{\rm\pi}=\frac{\rm 2\cdot\pi^{3}}{\rm 2\cdot\rm \pi\cdot \rm 3}=\frac{\rm \pi^2}{\rm 3} = \rm 3. 29. \hspace{0.5cm}\Rightarrow \hspace{0.5cm}\sigma_{\alpha}\hspace{0.15cm}\underline{=1.814}.$$
(4) Since the given section of the circle is exactly one quarter of the total circle area, the probability we are looking for is
- $${\rm Pr}(-π/4 ≤ α ≤ +π/4)\hspace{0.15cm}\underline{=25\%}.$$
(5) From simple geometrical ¨considerations (right-angled triangle, marked dark blue in the adjacent sketch) one obtains the equation of determination for the angle $\alpha_0$:
- $$\cos(\pi-\alpha_{\rm 0}) = \frac{R/ 2}{R}={\rm 1}/{\rm 2}\hspace{0.5cm}\Rightarrow\hspace{0.5cm}\rm\pi-\it\alpha_{\rm 0}=\frac{\rm\pi}{\rm 3} \hspace{0.2cm}\rm( 60^{\circ}).$$
- It follows $\alpha_0 = \pi/3\hspace{0.15cm}\underline{=2.094}.$
- This corresponds $\alpha_0 \hspace{0.15cm}\underline{=120^\circ}$.
(6) Correct is the suggested solution 3:
- The PDF $f_\alpha(\alpha)$ is f for a given angle $\alpha$ directly proportional to the distance $A$ between the antenna and the boundary line.
- For $\alpha = \pm 2\pi/3 = \pm 120^\circ$ against $A = R$, for $\alpha \pm \pi = \pm 180^\circ$ against $A = R/2$.
- In between the distance becomes successively smaller. This means: The PDF decreases towards the boundary.
- The decrease follows the following course:
- $$\it A=\frac{\it R/\rm 2}{\rm cos(\rm \pi-\it\alpha)}.$$
(7) The area $G$ can be calculated from the sum of the $240^\circ$–sector and the triangle formed by the vertices $\rm UVW$ :
- $$G=\frac{\rm 2}{\rm 3}\cdot \it R^{\rm 2}\cdot{\rm \pi} \ {\rm +} \ \frac{\it R}{\rm 2}\cdot \it R\cdot \rm sin(\rm 60^{\circ}) = \it R^{\rm 2}\cdot \rm\pi\cdot (\frac{\rm 2}{\rm 3}+\frac{\rm \sqrt{3}}{\rm 4\cdot\pi}).$$
- The probability we are looking for is given by the ratio of the areas $F$ and $G$ (see sketch):
- $$\rm Pr(\rm -\pi/4\le\it\alpha\le+\rm\pi/4)=\frac{\it F}{\it G}=\frac{1/4}{2/3+{\rm sin(60^{\circ})}/({\rm 2\pi})}=\frac{\rm 0.25}{\rm 0.805}\hspace{0.15cm}\underline{=\rm 31.1\%}.$$
- Although nothing has changed from point (4) at the area $F$ the probability now becomes larger by a factor $1/0.805 ≈ 1.242$ due to the smaller area $G$ .
(8) Since the overall PDF area is constantly equal $1$ and the PDF decreases at the boundaries, it must have a larger value in the range $|\alpha| < 2\pi/3$ than in (1).
- With the results from (1) and (7) holds:
- $$f_{\alpha}(\alpha = 0)=\frac{1/(2\pi)}{2/3+{\rm sin(\rm 60^{\circ})}/({\rm 2\pi})} = \frac{\rm 1}{{\rm 4\cdot\pi}/{\rm 3}+\rm sin(60^{\circ})}\hspace{0.15cm}\underline{\approx \rm 0.198}.$$
- Like the probability in (7) also simultaneously the PDF value in the range $|\alpha| < 2\pi/3$ increases by a factor $1.242$ as the coverage area becomes smaller.