24G FMCW Radar System And Algorithm Design Based On Novel Substrate Integrated Suspended Line

In recent years,with the development of the automotive industry,self-driving technology is gradually advancing and the status of advanced driver assistance systems(ADAS)is particularly important.Millimeter wave radar with all-weather and no blind spots have a role that other sensors cannot replace.The overall architecture design and algorithm research of 24 GHz FMCW radar based on a novel substrate integrated suspended Line is completed in this dissertation.In this dissertation,by analyzing the overall structure of the millimeter-wave radar We design the parameters of the radar.The center frequency of the radar is 25.25 GHz,the bandwidth is 300 MHz,and the frequency modulation period is 40 ms.The distance detection range of the radar is 5-100 meters,the range resolution is 1m,the speed measurement range is 0-40 m / s and the speed accuracy is 1m/s.We select the amplifier AMMP6425,the low noise amplifier AMP-6233,and the switch MC1084LC4 which are based on novel substrate integrated suspended Line.We select a six-port network,a Butler feed network,a four-element cavity back-projected dipole antenna which based on the novel substrate integrated suspended Line to design.We design two radar architectures,one is directly connected radar architecture and the other one is feed network radar architecture.We design the algorithm for this new radar,Radar IF signals are collected to obtain sampled signals.We select the fast Fourier transform algorithm for the sampled signal to obtain the sum and beat signals of the spectrum.Calculate the speed information and distance information by using the speed and distance formula.Radar angle measurement is completed through the time domain scan of the Butler feed network.We designed a radar scanning host computer interface based on QT for the radar frame.Finally,we use the signal source to verify the implementation of the radar algorithm,including ADC sampling verification,speed and distance verification.Based on the phase delay of the received target signal by the Butler feed network,we use the time-frequency information of the received signal and the classification characteristics of the convolutional neural network to design a radar precise angle positioning algorithm.The algorithm improves the angular resolution of the Butler feed network radar,and the resolution at this time is theoretically ± 1.5°.