| With the continuous development of unmanned driving technology,millimeter wave radar has been widely used.A lot of scholars at home and abroad have also studied and improved more related technologies in response to the ever-changing application scenarios.The function of the millimeter wave radar is to detect and extract the target motion information.Due to the increasingly complex driving environment including various external disturbances,a large amount of clutter and noise doped by the echo signal obscure the target information,which cause the radar cannot to detect the target object.In order to ensure the effectiveness of radar target detection,it is of great importance to improve the detection technology in complex environments.it is the key means to achieve radar target detection that CFAR detection technology,it can overcome the interference of various clutter in real environment and can improve the target detection probability efficiently while maintaining a low false alarm probability.However,the existing CFAR detection algorithm cannot ensure the detection adaptability when the location and quantity of clutter in the reference sliding window is uncertain.In this thesis,we therefore propose a new CFAR based on the improved Maximum Between-Cluster Variance from the perspective of classification.In addition,when performing CFAR detection on the beat signal,the noise in the signal is distributed in the reference unit,which produce the error of threshold estimation and affects the detection result.De-noising method of the radar signal based on the improved EMD and Kalman filtering is further proposed to reduce the impact of noise on the CFAR detection process.In this thesis,the target detection technology for the 77 GHz frequency modulated continuous wave(FMCW)radar is focused on research.The main research work and innovations of this thesis are as follows:(1)Introducing the research background of radar target detection algorithm and the research status of home and abroad.Analyzing the problems to be solved and making detailed planning for the research content and chapter arrangement of this thesis.(2)Brief overview of the radar signal processing process and target detection.The principle of empirical mode decomposition(EMD)and applications of de-noising are expounded.Finally,the principle and index of radar target detection are analyzed theoretically.(3)A new CFAR based on the improved Maximum Between-Cluster Variance is proposed.In order to eliminate the interference target unit adaptively and make full use of the uniform clutter unit information to estimate the background noise power which can improve the detection performance of the CFAR detector in multi-target interference environment,the mean square deviation and first-order difference concept are introduced into the Between-Cluster variance and the increment of the improved maximum inter-class variance is analyzed.Then,in order to estimate the background noise power and get the detect threshold,we make good use of the uniform clutter unit divided by the optimal threshold corresponding to the maximum increment.Simulation results show that the proposed algorithm has smaller CFAR loss which close to CA-CFAR in uniform environment and has more stable anti-interference ability in the multi-target jamming environment.(4)From the perspective of reducing the influence of noise on CFAR detection performance,the method of de-noising about beat signal is studied.This thesis proposes a joint de-noising method of radar signal based on improved EMD and Kalman filtering.We first analyze the change of the correlation coefficient between autocorrelation function about IMF component and noise-canceled beat signal in order to determine the modal boundary point.Then the de-noised signal can be obtained by processing the initial de-noised signal reconstructed by low frequency components below the modal boundary point with Kalman filter.Experiments show that the improved algorithm can find the modal boundary point more accurately and objectively,and the SNR of the output signal have been improved without losing the useful signal components.Finally,the de-noising method proposed in this chapter is combined with the improved CFAR detection algorithm to verify the change of radar target detection performance before and after de-noising. |
PDF Full Text Request