利用Python实现FMCW雷达的距离多普勒估计

参考代码：

#!/usr/bin/env python
# -*- coding:utf-8 -*-
# @Time  : 2020/9/24 21:29
# @Author: lg6
# @File  : fmcw_2dfft_multi.py


# coding=utf-8
import numpy as np
import matplotlib.pyplot as plt
import scipy as sp
from mpl_toolkits.mplot3d import Axes3D

# parameters setting

B = 135e6  # Sweep Bandwidth
T = 36.5e-6  # Sweep Time
N = 512  # Sample Length
L = 128  # Chirp Total
c = 3e8  # Speed of Light
f0 = 76.5e9  # Start Frequency
NumRangeFFT = 512  # Range FFT Length
NumDopplerFFT = 128  # Doppler FFT Length
rangeRes = c/2/B  # Range Resolution
velRes = c/2/f0/T/NumDopplerFFT  # Velocity Resolution
maxRange = rangeRes * NumRangeFFT  # Max Range
maxVel = velRes * NumDopplerFFT/2  # Max Velocity
tarR = [50, 90]  # Target Range
tarV = [3, 20]  # Target Velocity

# generate receive signal

S1 = np.zeros((L, N), dtype=complex)
for l in range(0, L):
    for n in range(0, N):
        S1[l][n] = np.exp(np.complex(0, 1) * 2 * np.pi * (((2 * B * (tarR[0] + tarV[0] * T * l))/(c * T) + (2 * f0 * tarV[0])/c) * (T/N) * n + (2 * f0 * (tarR[0] + tarV[0] * T * l))/c))

S2 = np.zeros((L, N), dtype=complex)
for l in range(0, L):
    for n in range(0, N):
        S2[l][n] = np.exp(np.complex(0, 1) * 2 * np.pi * (((2 * B * (tarR[1] + tarV[1] * T * l))/(c * T) + (2 * f0 * tarV[1])/c) * (T/N) * n + (2 * f0 * (tarR[1] + tarV[1] * T * l))/c))

sigReceive = S1 + S2

# range win processing

sigRangeWin = np.zeros((L, N), dtype=complex)
for l in range(0, L):
    sigRangeWin[l] = np.multiply(sigReceive[l], np.hamming(N).T)

# range fft processing

sigRangeFFT = np.zeros((L, N), dtype=complex)
for l in range(0, L):
    sigRangeFFT[l] = np.fft.fft(sigRangeWin[l], NumRangeFFT)

# doppler win processing

sigDopplerWin = np.zeros((L, N), dtype=complex)
for n in range(0, N):
    sigDopplerWin[:, n] = np.multiply(sigRangeFFT[:, n], np.hamming(L).T)

# doppler fft processing

sigDopplerFFT = np.zeros((L, N), dtype=complex)
for n in range(0, N):
    sigDopplerFFT[:, n] = np.fft.fftshift(np.fft.fft(sigDopplerWin[:, n], NumDopplerFFT))


fig = plt.figure()
ax = Axes3D(fig)


x = np.arange(0, NumRangeFFT*rangeRes, rangeRes)
y = np.arange((-NumDopplerFFT/2)*velRes, (NumDopplerFFT/2)*velRes, velRes)
# x = np.arange(NumRangeFFT)
# y = np.arange(NumDopplerFFT)
# print(len(x))
# print(len(y))
X, Y = np.meshgrid(x, y)
Z = np.abs(sigDopplerFFT)
ax.plot_surface(X, Y, Z,
                rstride=1,  # rstride（row）指定行的跨度
                cstride=1,  # cstride(column)指定列的跨度
                cmap=plt.get_cmap('rainbow'))  # 设置颜色映射

ax.invert_xaxis()  #x轴反向

plt.show()

 

仿真结果：

 

 

参考资料：

干货 | 利用MATLAB实现FMCW雷达的距离多普勒估计

https://mp.weixin.qq.com/s?__biz=MzI2NzE1MTU3OQ==&mid=2649214285&idx=1&sn=241742b17b557c433ac7f5010758cd0f&chksm=f2905cf9c5e7d5efc16e84cab389ac24c5561a73d27fb57ca4d0bf72004f19af92b013fbd33b&scene=21#wechat_redirect

干货| 利用Python实现FMCW雷达的距离多普勒估计

https://mp.weixin.qq.com/s/X8uYol6cWoWAX6aUeR7S2A

公式推导建议：

 Intro to mmWave Sensing : FMCW Radars 

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