Research Of Technology Of Intermediate Frequency Signal Processing For FMCW Radar Ranging

FMCW radar can measure the distance and speed of a static or moving object, which is now widely used in the highway speed monitoring system, anti-collision radar and other fields. While driving the super heavy truck transporting in a large coal mine station, the driver often travels around in a blind area. So there is existing lots of potential accidents. This paper chooses to install FMCW radar on the truck in order to measure the distance between the obstacle and car body in real time and insure the safety of the personal and property. In order to extract the high accuracy distance information from the radar difference frequency signal, it is necessary to make a processing to filter out the noise and estimate the frequency of difference frequency signal. The research in this paper will focus on the two aspects.In the first, this paper discusses the structure of FMCW radar and the principle of radar ranging. The pre-processing circuit of difference frequency signal is designed in this paper which is including analog filters, adjustable gain amplifies.And a DSP development board based on TMS320F28335is designed for signal processing. In the following digital signal processing, difference frequency signal is analyzed and tested by two signal denoising and filtering methods which is based on its own characteristics.The paper designes a FIR digital filter in order to separate and filtering the interference signal that exists the fixed band from difference frequency signal.The features of denoising by FIR filter is simple, easy to make a simulation and hardware implementation; For the problem that wide-band random noise can’t use a fixed cut-off frequency filter to separate, this paper introduces a EMD denoising method based on the energy characteristics of autocorrelation function and puts forward a new adaptive method to decide the point of useful signal dominant mode and noise dominant mode. Finally, the simulation experiment of the actual difference frequency signal was carried out. Based on the analysis of the energy spectrum estimation theory in the energy centrobaric correction method and Chirp-Z transform, the traditional FFT algorithm is modified to improve the spectrum estimation precision. This paper introduces a target distance estimation algorithm that uses FFT+Chirp-Z transform based on the energy centrobaric characteristics.The algorithm compared with the conventional partial spectrum zoom analysis algorithm can halve the Chirp-Z transform points and further improve the estimation accuracy of the frequency.By this method,it is more conducive to realize the algorithm in DSP hardware.The paper collected several sets of difference frequency signal in the distance of1～30m by MP424card and made an overall hardware and software test. The experimental results show the effectiveness of EMD denoising algorithm and verify the rapidity of improved FFT+Chirp-Z transform algorithm.Compared with traditional FFT algorithm, it is faster and accurate in measuring distance.