A Multi-target Detection Method Based On Signal Power In FMCW Radar

As Chinese economy developing,people's living conditions have been greatly improved.Cars that were high-end products decades ago have now entered thousands of households.However,although cars have brought great convenience to people's life,the rapidly increasing number of cars has also brought inevitable traffic safety problems.In order to reduce the loss caused by traffic accidents,the research of vehicle-mounted anti-collision radar has become a new research hotspot in recent years.Frequency modulated continuous wave radar stands out from many sensors because of its advantages of simple modulation,weather resistance and high resolution.This paper focuses on the research of vehicle-mounted anti-collision radar,aiming at the problem of ghost targets generated by traditional triangular waveform in multi-target detection,proposes a multi-target detection method based on signal energy,which enables triangular wave to complete multi-target recognition without generating ghost targets.Firstly,the research background and significance of vehicle-mounted anti-collision radar are described,and the research status of vehicle-mounted anti-collision radar in China and other contries is introduced.Then the basic waveforms commonly used in frequency modulated continuous wave radar,the improved waveforms and the new signal processing methods are briefly introduced.The structure diagram of the paper is given at the end of first chapter.Secondly,the basic structure of radar is introduced.The reason why continuous wave radar is more suitable for early warning and anti-collision task is explained.Then,the signal processing flow of continuous wave radar and the principle of basic triangular waveform are described,and the advantages and problems of triangular waveform are pointed out.Then,three existing improved waveforms of triangular waveform are introduced.In view of unnecessary calculation and insufficient use of existing information in the processes of the three algorithms,the steps of the algorithms are adjusted and optimized.The pairing calculation is greatly reduced by using the big difference between the two frequencies which are needed.By adjusting the sequence of eliminating ghost targets and calculating real ones,the times of multiplication calculation are greatly reduced.The improved algorithm flow chart is given and the performance of the three algorithms is simulated.Finally,aiming at the problems of generating ghost targets in multi-target detection of basic triangle waveform,large calculation amount of improved waveform,complex waveform and so on,a multi-target detection method based on signal energy using triangle waveform is proposed.Since the echo energy varies according to the target distance,the spectrum amplitude outputs by the Fourier transform can be used to assist multi-target recognition.However,due to the existence of spectrum leakage,there may be a large error between the observed spectrum amplitude and the actual value,so it cannot be used for multi-target detection directly.In this paper,an existing non-iterative method of estimating spectrum fractional part is used to estimate fractional part and to modify the spectrum amplitude.The error between the corrected spectrum amplitude and the true value is about 3%,which indicates the accuracy of the amplitude correction.Then the feasibility and effectiveness of the proposed algorithm are proved by theoretical derivation.The simulation results show that the proposed algorithm not only enables the basic triangular wave to recognize the multi-objective situation,but also has lower calculational complexity and higher estimation accuracy than the existing improved waveform method under same conditions.