TI mmWave radar sensors Tutorial 笔记 | Module 1: Range Estimation

本系列为TI(Texas Instruments) mmWave radar sensors 系列视频公开课 Module 1: Range Estimation 的学习笔记。

• 视频网址: https://training.ti.com/intro-mmwave-sensing-fmcw-radars-module-1-range-estimation?context=1128486-1139153-1128542

• 关注 下面的公众号，回复“ TI毫米波 ”，即可获取 本系列完整的pdf笔记文件~

---

内容在CSDN和微信公众号同步更新

• CSDN博客
• 微信公众号

• Markdown源文件暂未开源
• 笔记难免存在问题，欢迎联系指正

当前更新列表：

• MIT 矩阵方法
• 《合成孔径雷达成像原理》
• TI mm wave Tutorial

---

---

1.1 Introduction to mmwave Sensing – Module 1: Range Estimation

• FMCW Radars

  • F requency M odulated C ontinuous W aves

  • very popular in both automotive and industrial segments

  • basically measures the range , velocity and angle of arrival of objects

    ✅ 接下来介绍 each of these demensions of sensing:

    🚩 range (module 1) $\Rightarrow$ 🚩 velocity $\Rightarrow$ 🚩 angle estimation(module 5)

• Contents in this module :

  • Basics of FMCW radar operation
  • Using the radar to measure range of multiple objects in front of radar
  • Concept of IF signal and IF Bandwidth
  • Range Resolution

• Aim: Try to answer the following questions

  • 单目标测距 ：How does the radar estimate the range of an object?
  • 多目标测距 ：What if there are multiple objects?
  • 分辨力问题 ：How close can two objects get and still be resolved as two objects?
  • 最远探测距离 ：What determines the furthest distance a radar can see?

What is a chirp?

• chirp
  • a sinusoid whose frequency increases linearly with time
  • 如下图 A-t plot (A: Amplitude)

• Start at frequency $f_c$ ; End at frequency $f_{c+B}$

  • B : the bandwidth of the chirp

  • $\Rightarrow$ a chirp is a continuous wave of whose frequency is linearly modulated (FMCW)

  • its frequency-time (f-t) plot

  • S (slope of the chirp): the rate at which the chirp ramps up (e.g., sweeping a bandwidth of 4GHz within 40us $\Rightarrow$ a Slope of 100MHz/us)

    🚩 B and S are important parameters which define system performance

Now that we know what a chirp is

• next: how an FMCW radar works
  

A 1TX-1RX FMCW Radar

• A simplified block diagram of an FMCW radar
  • with a single TX and a single RX

• The radar operates as follows:
  • 1 A synthesizer (synth) generates a chirp
  • 2 The chirp is transmitted by the TX antenna
  • 3 The chirp is reflected off an object and the reflected chirp is received at the RX antenna
  • 4 The RX signal and TX signal are ‘mixed’ and the resulting signal is called an “IF signal”

IF: intermediate frequency

• What is a Mixer ?

  • A mixer: a 3 port device with 2 inputs and 1 output

  • 2 inputs : $x_1=sin({\omega }_1+{\varphi }_1)$ and $x_2=sin({\omega }_2+{\varphi }_2)$

  • 1 output : $x_3=x_{out}=sin[({\omega }_1-{\omega }_2)t+({\varphi }_1-{\varphi }_2)]$ , with two properties:

    🚩 Instantaneous Instantaneous 瞬时 frequency equal to the difference of the instantaneous frequencies of the two input sinusoids

    🚩 Phase equal to the difference of the phase of the two input sinusoids

推导：积化和差公式

• 再利用IF filter 去掉 $({\omega }_1+{\omega }_2)$, 只保留$({\omega }_1-{\omega }_2)$

The IF signal

• The TX-signal and the RX-signal reflected from an object (上图)
  • The RX is just a delayed version of TX signal
  • $\tau$ : the round-trip time between the radar and the object
• TX chirp 和 RX chirp 之间的frequency difference即为 IF (intermediate frequency )
  • $\Rightarrow$ A single object produces an IF signal – a sonstant frequency tone
  • the frequency of IF: $S\tau =S\times 2d/c=S2d/c$

结论: A single object in front of the radar produces an IF signal with a constant frequency of $S2d/c$

• Note 1 : The IF signal is valid from time $\tau$ (the time the reflected signal is received)
  • ADC模数转换时需要注意开始时间
  • End time: $T_c$
• Note 2 : $\tau <<T_c$
  • $\tau$ is typically a small fraction of the total chirp time
  • 300米探测距离, $T_c=40us$的 radar: $\tau /T_c=5\%$

Fourier Transforms: A quich review

• Fourier transform :

  • very important in FMCW radar!
  • used in range, velocity, and angle estimation

• Reivew:

  • FT converts a signal from time domain $\Rightarrow$ frequency domain
  • A sinusoid in the time domain produces a single peak in the frequency domain

关于 频率混淆 $\Rightarrow$ 决定 距离分辨率 ：

• Within the observation window T below, 频率差仅 $1/(2T)$ $\Rightarrow$ not sufficient to resolve the tones in the frequency spectrum

• Doubling the obeservation windows $\Rightarrow$ 观察窗口内 results in a difference of 1 cycle

结论：Longer the observation period $\Rightarrow$ better the resolution

• observation window T: 能够区分频率差大于$1/T$ 的 frequency components

Multiple objects in front of the radar

• Multiple objects in front of the radar
  • $\Rightarrow$ multiple reflected chirps at the RX antenna
  • $f=S2d/c$ proportional to the range $\Rightarrow$ 距离越远，$\tau$越大，IF频率越大

• A frequency spectrum of the IF signal will reveal multiple tones
  • 且 频域 与 object frequency 成正比

Range Resolution in a radar

• Range resolution refers to the ability to resolve two closely spaced objects
  • T window : $\Rightarrow$ $> 1/T $ 的频率差
• In this slide, the two objects are too close that they show up as a single peak in the frequency spectrum
  • 从A-T图中 也可以看出 the difference between frequencies is too small

$\Rightarrow$

• Solutions :

  • extend the observation window T $\Rightarrow$ until ${\Delta }_F>1/T$

    ✅ i.e., increase $T_c$ $\Rightarrow$ 增加带宽 $B=T_c\times S$

A clue: possibly a larger bandwidth $B$ corresponds to a better range resolution

• 下面进行说明

Range Resolution in a radar

• Recall that:

  • IF frequency : $f=S2d/c$
  • Two tones can be resolved in frequency: $\Delta f>1/T$

• Next: derive an equation for the range resoltion of the radar

  • 距离分辨率依赖于那些参数？Chirp Duration, Bandwidth, Slop?

• Derive:

  • $\Delta f=\frac{S2\Delta d}{c}>\frac{1}{T_c}$ $\Rightarrow$
  • $\frac{\mathrm{S}2\Delta \mathrm{d}}{\mathrm{c}}>\frac{1}{{\mathrm{T}}_{\mathrm{C}}}$ $\Rightarrow$ $\Delta \mathrm{d}>\frac{\mathrm{c}}{2{\mathrm{ST}}_{\mathrm{C}}}$
  • $\Rightarrow$ $\Delta d>\frac{c}{2B}$

结论 : The range Resolution ($d_{res}$) depends only on the Bandwidth swept by the chirp :

• $d_{res}=\frac{c}{2B}$

---

• Question : Which of these two chirps gives a better range-resolution?
  • What is the intuition behind this reuslt?

• Ans :

  • Same B $\Rightarrow$ Same range-resolution

• 然而， intuitively ，$T_c$的大小也应该会造成一定的影响

  • consider the properties of Fourier Transforms
  • (long time $\Rightarrow$ narrow frequency $\Rightarrow$ Higher resolution
  • How to resolve this contradiction ?
  • 见下文

---

Digitizing the IF signal

• The bandwidth of interest of the IF signal depends on the desired maximum distance:
  • ${\mathrm{f}}_{{\mathrm{IF}}_{-}\max }=\frac{\mathrm{S}2{\mathrm{d}}_{\max }}{\mathrm{c}}$
  • and then, determine the cut off frequency of the LPF
• IF signal --> LPF + ADC (digitized) --> further processing on a DSP
• 因此，ADC采样率(F_s)应该满足：$F_s\ge \frac{S2d_{max}}{c}$
  • (We assume a complex baseband signal $\Rightarrow$ half the Nyquist rate of a real signal)

结论 : An ADC sampling rate of $F_s$ limites the maximum range of the radar to $d_{max}=\frac{F_{s}c}{2S}$

• 相同采样率时，S越小($T_c$越长)，最大探测距离越大 (代价是$T_c$变长，相当于 以时间换空间 )

---

• Now, we can Revisit our ealier example
  • Chirp A and B has the same bandwidth and range resolution
  • 但是，For the same max detection range $d_{max}$, A needs smaller sampling rate $F_s$ but more time
  • So here is a trade-off

Summary

• The process:

  • totally, 5 steps

  • 1 generate a chirp

  • 2 transimit and receive the reflected chirp

  • 3 IF signal and LPF

  • 4 ADC

  • 5 An FFT is performed on the ADC data

    🚩 $f_{IF}=\frac{S2d}{c}$

    🚩 因此，直接在时间信号上做FFT，得到频域

    🚩 The location of peaks in 频域 $\Rightarrow$ directly correspond to the range

    🚩 This FFT is called a “range-FFT” $\Rightarrow$ 因为it results objects in range

Key concepts / formulas

• $f_{IF}=\frac{S2d}{c}$
• $d_{res}=\frac{c}{2B}$
• $d_{max}=\frac{F_{s}c}{S}$
• 两个带宽 : Chirp Bandwidth VS IF bandwidth
  • Large chirp bandwidth $\Rightarrow$ better range resolution
  • Large IF Bandwidth $\Rightarrow$ faster chiprs, 更小的maximum distance

问题引出: velocity

• Two objects equidistant from the radar
  • How will the range-FFT look like?

• Ans:
  • 如果它们 速度 不同，那就可以separated out by further signal processing
  • See in the next module