提出一种用于汽车雷达防撞技术的全频目标识别(FSTR)算法,该方案采用24 GHz毫米波雷达,在FPGA平台上进行信号处理.通过python软件实现算法并进行仿真,仿真结果和FPGA实际运行的结果一致.该方案的雷达扫频信号为100 Hz的锯齿波,采样频率为480 kHz,离散傅里叶变换(DFT)点数为4096点,每一帧数据的时间间隔为10 ms,满足车辆行驶实时性的要求.通过对道路行驶测得的数据比较,相比主流的恒虚警率(CFAR)算法,本算法抗干扰能力更强,精确度更高.%A full spectrum target recognition(FSTR)algorithm used for collision avoidance technology of automotive radar is proposed. A 24 GHz millimeter-wave radar is used in the scheme to process the signal on FPGA platform. The FSTR algo-rithm is realized with the python software,and its simulation result is consistent with the actual running result processed by FP-GA. The items of the scheme can meet the real-time requirement of vehicle travel,which includes that the radar sweet-frequency signal is the 100 Hz sawtooth wave,the sampling frequency is 480 kHz,the discrete Fourier transform(DFT)count is 4096, and the time interval of each frame data is 10 ms. The comparison result of the data measured in vehicle travelling shows that, in comparison with the constant false-alarm rate (CFAR) algorithm,the FSTR algorithm has stronger anti-interference ability and higher accuracy.
展开▼
