###### tags: `課程筆記`
# 訊號與系統 Signal and System
> 是指對訊號表示、轉換、運算等進行處理的過程,就是要把記錄在某種媒體上的訊號進行處理,以便抽取出有用資訊的過程,它是對訊號進行提取、轉換、分析、綜合等處理過程的統稱。
[TOC]
- <span style="background-color: #e2f8f0; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;"></span>
- <span style="background-color: #c8d8df; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;"></span>
- <span style="background-color: #ffd7c1; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;"></span>
## 1. Basic Concepts
### 1.3 Classifications of Signals
- Continuous vs Discrete
- continuous time signal $x(t)$
- ![](https://i.imgur.com/tT6VgS2.png)
- discrete time signal $x[n]$
- 在特定時間上才有值,通常都為固定在 continuous-time 做 signal sampling 得出
- signal is defined at particular instants(立即) of time
- 值仍是 continuous
- 需要做 quantization 完才會是 discrete signal 表值與時間皆為 discrete
- ![](https://i.imgur.com/e8lNgHH.png)
- discrete time 表示法
- $x[k] = x[kT]$
- $x[k]$ 表第 k 個取樣
- $x[kT]$ 表第 kT 時間取樣
- $x[kT]$ 的表示法比較好
- 可以直接看出時間點
- 未來也可以與其他數據結合搭配,所以時間單位再結合時,也需要校正
- Periodic and Nonperiodic Signal
- periodic continuous $x(t) = x(t+nT)$
- ![](https://i.imgur.com/wwpx8OK.png)
- periodic discrete $x[n] = x[n+N]$
- ![](https://i.imgur.com/Bw1eSTN.png)
- <span style="background-color: #e2f8f0; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">【補充】Euler’s Identities</span>
- $e^{j\theta} = cos\theta + jsin\theta$
- $e^{-j\theta} = cos\theta - jsin\theta$
- ![](https://i.imgur.com/bS3TNQI.png)
- ![](https://i.imgur.com/BqUSNvD.png)
- An analog signal vs A digital signal
- An analog is a continuous time signal in which the variation with time is analogous (or proportional) 時間與值都是連續的
- is a discrete time signal that can have a **finite number** of values (usually binary).時間與值都不是連續的
- 需要 sampling and quantization
- Energy vs Power Signals
- ![](https://i.imgur.com/JERkkHR.png)
- A signal $x(t)$ or $x[n]$ is an **energy signal** if and only if `代E公式會介於0到無窮大` and `代P公式會等於 0`
- A signal $x(t)$ or $x[n]$ is a **power signal** if and only if `代P公式會介於0到無窮大` and`代E公式會等於無窮大`
- <span style="background-color: #ffd7c1; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">所以當題目給一個 $x(t)$ 或是 $x[n]$ 時,就直接帶 $E$ 或 $P$ 的公式,看哪一個會介於0到無窮大之間</span>
- Even vs Odd smmetry
- even:$x(t) = x(-t)$
- odd:$x(t) = -x(-t)$
- 每一個 signal 都可以被分為 even part 跟 odd part
- $x(t) = x_e(t) + x_o(t)$
- $x_e(t) = \frac{1}{2}[x(t) + x(-t)]$(把odd part 部分給消掉)
- $x_o(t) = \frac{1}{2}[x(t) - x(-t)]$(把even part 部分給消掉)
### 1.4 Basic continuous-time signals
- **Unit impulse function(delta function)**
- ![](https://i.imgur.com/kxacMTT.png)
- $\int_{0-}^{0+} \delta(t) \,dt = 1$
- sampling
- ![](https://i.imgur.com/ICQQwCV.png)
- **Unit step function**
- ![](https://i.imgur.com/glbaOYT.png)
- **Unit ramp function**
- ![](https://i.imgur.com/m7ixblY.png)
- <span style="background-color: #c8d8df; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">3 種函數關係</span>
- $r(t)$ 微分變 $u(t)$, $u(t)$ 微分變 $\delta(t)$
- $\delta(t)$ 積分變 $u(t)$, $u(t)$ 積分變 $r(t)$
- 其他延伸的 function
- <span style="background-color: #e2f8f0; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">**Rectangular Pulse Function**![](https://i.imgur.com/8i9mFVt.png)</span>
- **Triangular Pulse Function**
- ![](https://i.imgur.com/Yzc7ZcP.png)
### 1.5 Basic discrete-time signals
- **Unit impulse sequence**
- ![](https://i.imgur.com/FEXZ7TO.png)
- **Unit step sequence**
- ![](https://i.imgur.com/IDIcnrh.png)
- **Unit ramp sequence**
- ![](https://i.imgur.com/8YMIW2n.png)
- 三者相對關係
- ![](https://i.imgur.com/QS80vza.png)
- ![](https://i.imgur.com/hz5nJbq.png)
- 其他函數圖形
- **Sinusoidal sequence**
- ![](https://i.imgur.com/fSA0Eum.png)
- **Exponential Sequence**
- $x[n] = Ae^{-anT} = A\alpha^n, \alpha = e^{-aT}$
- ![](https://i.imgur.com/uhHaVSt.png)
### 1.6 Basic operations on signals
- **Time Reversal**
- ![](https://i.imgur.com/8o9aeby.png)(此圖有誤)
- **Time Scaling**
- ![](https://i.imgur.com/y7EPJkJ.png)
- **Time Shifting**
- ![](https://i.imgur.com/Q87hmFp.png)
- **Amplitude Transformations**
- ![](https://i.imgur.com/2ycEkPU.png)
- ![](https://i.imgur.com/xOSVgfR.png)
### 1.7 Classification of systems
- ***Continuous Time and Discrete Time Systems***
- ![](https://i.imgur.com/5DZX0MA.png)
- ***Causal and Noncausal Systems***
- A causal system is one whose present response does **not depend on the future values** of the input
- ***Linear and Nonlinear Systems***
- Homogeneity + Additivity
- ![](https://i.imgur.com/fvMNXJq.png)
- ***Time Varying and Time Invariant Systems***
- vary:to change or cause something to change in amount or level, especially from one occasion to another
- ![](https://i.imgur.com/XcPJ1QG.png)
- time-invariant
- ![](https://i.imgur.com/kzRcESm.png)
- time-varying
- ![](https://i.imgur.com/1aF9MzM.png)
- ***Systems with and without Memory***
- When the output of a system **depends on the past or future input**, the system is said to have a memory.
## 2. Convolution
- [But what is a convolution?](https://www.youtube.com/watch?v=KuXjwB4LzSA)
### 2.2 Impulse response
> The impulse response to an LTI (linear, time invariant) system is **`the output of the system to a unit impulse function`**.
- Impulse response(green part)
- ![](https://i.imgur.com/uhD8b8c.png)
### 2.3 Convolution integral
- The convolution of two signals $x(t)$ and $h(t)$ is usually written in terms of the operator $*$ as
- ![](https://i.imgur.com/93R2cLh.png)
- <span style="background-color: #ffd7c1; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">我們 input signal $x(t)$ 與 impulse function $\delta(t)$,經過 LTI 系統的 transform,我們可以得到一個 impulse response $h(t)$,接著 $x(t)$ 再與 $h(t)$ 做 convolution 後,就可以得到 output $y(t)$</span>
- 若 $x(t) = 0$ for $t < 0$, 且 the system is causal, $h(t) = 0$ for $t < 0$(此處的 $\tau$ 為變數)
- 則![](https://i.imgur.com/aTrR7we.png)
- 或是也可寫成 $y(t) = x(t)*h(t) = \int_{0}^{t}x(t-\tau)h(\tau)d\tau$
### 2.4 Graphical convolution
- 使用到前方提過的函數技巧
![](https://i.imgur.com/KOtJ9m4.png)
- 需要分段討論的範例
- ![](https://i.imgur.com/T2eyDqx.png)
- 根據定義可列
- ![](https://i.imgur.com/4CMS8ZT.png)
### 2.5 Block diagram representation for continuous
![](https://i.imgur.com/vg5LsPg.png)
### 2.6 Discrete-time convolution
- 定義
- `2.2、2.3 的離散版本`
- ![](https://i.imgur.com/5auesKV.png)
- **範例** Find $y[n] = x[n]*h[n]$ ![](https://i.imgur.com/6z8Lilc.png)
- **`法一分析`** ![](https://i.imgur.com/Sv1H1JX.png)
- **`法二圖解`** ![](https://i.imgur.com/hvaWvoe.png)
- <span style="background-color: #e2f8f0; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">意義上為,將 $h(\tau)$ 先做 mirror(對y軸對稱),再分別向右平移 $0\sim t$、每次都與 $x(\tau)$ 相乘,最後再合成。
</span>
- 推廣在 $matlab$ 的運算方式![](https://i.imgur.com/8B6oSyZ.png)
### 2.7 Block diagram realization for discrete
![](https://i.imgur.com/aH2GEom.png)
- **範例** Let $x[n] = \{3, 0, 2, 6\}$ and $y[n] = \{6, 12, 25, 20, 38, 42\}$. Find $h[n]$
- 法一長除法:![](https://i.imgur.com/fdn1g2h.png)![](https://i.imgur.com/doQOWDs.png)
- 法二遞迴:![](https://i.imgur.com/3VPhQne.png)
## 3. The Laplace Transform
### 3.1 Introduction
- Laplace Transform 是將 linear system 轉換為 **`frequency-domain`** 的表達方式
- 能夠將常微分方程式(ordinary differential equations)轉為代數方程(algebra equations),以利於運算
- 讓 convolution 成為簡單的乘法
- 在 continuous-time LTI (Linear Time-invariant) system 產生 transfer function
### 3.2 Definition of laplace transform
- Laplace Transform:![](https://i.imgur.com/Dv6L0cP.png)
- Inverse Laplace Transform:![](https://i.imgur.com/FVRHlus.png)
- Laplace transformable:![](https://i.imgur.com/0eKJP0t.png)
- Region of Convergence(ROC):Laplace transform 就是會收斂的範圍 $Re(s) = σ >σ_c$![](https://i.imgur.com/nGTGGYn.png)
- example:![](https://i.imgur.com/MLD07Tt.png)![](https://i.imgur.com/27mzcrB.png)![](https://i.imgur.com/pxDEEGh.png)![](https://i.imgur.com/rIicIeL.png)
### 3.3 Properties of the laplace transform
- ***Linearity***:![](https://i.imgur.com/klec6Cg.png)
- ![](https://i.imgur.com/0eLU9zm.png)
- ***Scaling***:![](https://i.imgur.com/KLsjMi5.png)
- ![](https://i.imgur.com/5ocdoy8.png)
- ***Time Shifting***:![](https://i.imgur.com/B99c37h.png)
- ![](https://i.imgur.com/mV6TsNe.png)
- ***Frequency Shifting***:![](https://i.imgur.com/GC6Lczw.png)
- ![](https://i.imgur.com/56e2J4X.png)
- ***Time Differentiation***:![](https://i.imgur.com/uvyQIBZ.png)
- ![](https://i.imgur.com/wFVRE2y.png)
- ***Time Convolution***:![](https://i.imgur.com/oTnTLn4.png)
- ![](https://i.imgur.com/Ed82dE1.png)
- ***Time Integration***:![](https://i.imgur.com/eX79Cve.png)
- ![](https://i.imgur.com/dtqhXzm.png)
- **Frequency Differentiation**:![](https://i.imgur.com/x7mGx4i.png)
- ![](https://i.imgur.com/bElyiQt.png)
- **Initial and Final values**:已知![](https://i.imgur.com/uvyQIBZ.png),當![](https://i.imgur.com/z733Wow.png)![](https://i.imgur.com/1hbh2QF.png)
- all Properties:
- ![](https://i.imgur.com/dJEslYP.png)
- ![](https://i.imgur.com/j2AAxrL.png)
### 3.4 The inverse laplace transform
- ***Definition***:![](https://i.imgur.com/kCAjv4p.png)
- ***Simple Poles***:
- 已知:<span style="background-color: #c8d8df; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">我們已知![](https://i.imgur.com/pxDEEGh.png),故在 Simple poles 可以直接從 $X(s)$ 回推 $x(t)$</span>
- formate:![](https://i.imgur.com/cUbk2t0.png)
- ***Repeated Poles***
- 已知:<span style="background-color: #ffd7c1; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">因為 $(s+p)^n$ 出現,因此需要 **`微分`** 來找 $k$,且![](https://i.imgur.com/ZtJAEUu.png)(可以由 Frequency Differentiation 推導)
</span>
- format:![](https://i.imgur.com/KXc6jnj.png)
- 補充:當看到 $L^{-1}[\frac{1}{s^2}] = t$ 也是用此 pole
- ***Complex Poles***:
- 已知:<span style="background-color: #e2f8f0; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;"> ***Frequency Shifting*** ![](https://i.imgur.com/UaEfH5N.png)
、![](https://i.imgur.com/UcX0v4P.png)故 Complex Poles 可以應用</span>
- formate:**先從分母找 $\alpha$、$\beta$,再調整分子看 A B**![](https://i.imgur.com/cw3KFQM.png)
- 延伸:$sin$ 跟 $cos$ 可以合併![](https://i.imgur.com/XIJsLa3.png)
### 3.5 Transfer function
- 定義:<span style="background-color: #ffd7c1; display: block; padding: 2% 2% 0.5% 2%; border-radius: 15px;">$H(s)$為輸出結果 $Y(s)$ 與輸入結果 $X(s)$ 的比值,如此一來,再利用 Inverse Laplace Transfer,就可以輕鬆找到 $h(t)$</span>
- ![](https://i.imgur.com/OXWPVfj.png)
- ***Cascade connection***:![](https://i.imgur.com/nPuj0EC.png)
- ***Parrallel interconnection***:![](https://i.imgur.com/3pc3rYq.png)
- ***Feedback interconnection***:![](https://i.imgur.com/Az2MVQo.png)
### 3.6 Integro-Differential equations
![](https://i.imgur.com/z22FJPc.png)
## 4. Fourier Series
## 5. Fourier Transform
## 6. Discrete Fourier Transform
## 7. z-Transform
- 我們需要 z-transform 的原因是 Fourier transform 沒辦法 converge for all sequences
- 傅立葉轉換![傅立葉轉換](https://hackmd.io/_uploads/B1V3trY10.png)
- z-transform![z-transform](https://hackmd.io/_uploads/B1ipKrt1A.png)![截圖 2024-04-02 下午5.04.01](https://hackmd.io/_uploads/SkpZ5HK1C.png)
- one-sided or unilateral z-transform![截圖 2024-04-02 下午5.04.43](https://hackmd.io/_uploads/rkPEcSYk0.png)