| Nowadays,FMCW millimeter wave radars are commonly applied in automotive radar systems,such that the probability of mutual interference will be increased.As a result,the detection performance of radar system will be seriously degraded.When a signal transmitted by another radar has been received by the victim radar directly or indirectly,the accracy of target detection is influenced for the rising noise floor,which is the cause of misdetection resulting in car accidents.Therefore,the problem of mutual interference between FMCW radar systems has seen increased importance.At present,methods to deal with interference can be categorized into two types,which are proactive method and reactive method.The proactive method aims to lower the probability of interference,and the reactive method aims to reduce the impact of interference after the radar signal was interfered.Among the reactive methods,the method based on change-point detection and signal reconstruction is effective.In this method,the positions of the interferers in time region can be detected by the change-point detection algorithm,and the desired signals can be reconstructed by the signal reconstruction algorithm.Therefore,the researches on the method of interference suppression based on change-point detection and signal reconstruction is proposed in the thesis.Moreover,the accuracy of DOA estimation by beaformer of MIMO radar will be influenced after interference has occurred.As a result,a method based on change-point detection algorithm for DOA estimation with interference is also proposed in this thesis.Firstly,the methods to deal with interference are researched.Among the proactive methods,Frequency Division Multiplex Access(FDMA)philosophy,Time Division Multiplex Access(TDMA)philosophy and adding Offset method are introduced.Among the reactive methods,according to two different mixing signal cases,we set up the corresponding model for the situation,in which inteferers occuer.Then,we conclude that a suitable algorithm should be applied to the interference suppression in the beat signal with interference,after analyzing the signal model.Secondly,an algorithm,called PELT-KCN,is proposed.When pruned exact linear time(PELT)algorithm is used to detect the positions of interferences in time domain for mutual interference suppression algorithm based on signal restoration in FMCW radar,the detection results are not exact for the uncertain optimal penalty factor of PELT.Therefore,this thesis proposes a PELT-KCN algorithm for mutual interference suppression in FMCW radar.First,weighted-envelope normalization(WEN)is applied to find the envelope of time-domain signal of radar.Second,changepoints number can be obtained by analyzing the amplitude of the envelope.Third,according to the formula,the optimal penalty factor is obtained by using changepoints number.Then PELT algorithm is used to detect the time domain position of the interference based on the optimal penalty factor.Finally,the signal without interference is restored by an autoregressive model.Simulation and experiment results show that the proposed algorithm can mitigate the effect of the interference and improve the signal-to-noise ratio of detected targets in the case of single or multiple interferences.Thirdly,a DOA estimation for interference suppression based on change-point detection is proposed in this thesis,for the problem of low SNR and low detection accuracy in DOA estimation of beaformers caused by mutual interference.First,the interferers of the beat signals from arrays are suppressed by the suppression method,in which the indexes of the interfers in beat signal sequence are detected by change-point detection,and the signals distorted by the interfers are restored according to other normal part.Then the processed signals are transferred to the beaformer for DOA estimation.The results of the simulation and field experiment have verified that the proposed method can effectively increase the SINR corresponding to the targets in the space spectrum produced by the beamformers and increase the accuracy of DOA estimation. |
PDF Full Text Request