Design And Implementation Of Super Resolution Real-time Direction Finding System

Direction of Arrival(DOA),which is to determine the direction of the signal source through spatial spectrum estimation,is one of the key research directions in the field of modern array signal processing.The multiple signal classification algorithm is a classic algorithm for direction of arrival estimation,which has the advantages of high accuracy and super resolution.However,due to the large amount of calculation,the complexity of hardware implementation and the difficulty of real-time performance to meet actual needs,the hardware implementation faces huge challenges.Therefore,it is still of great significance to study the direction of arrival estimation in hardware.This thesis mainly studies the realization of super-resolution direction finding system based on MUSIC algorithm.The direction of arrival estimation is mainly for uniform linear arrays.The system scheme adopts the MUSIC algorithm in spatial spectrum estimation.The MUSIC algorithm is mainly divided into four parts in realization: covariance calculation,eigen decomposition,noise subspace acquisition and spectral peak search.The hardware circuit adopts the common structure(FPGA + DSP).The FPGA is used to perform preprocessing for the converted signal and controlling interface.Finally,the multi-core DSP is responsible for the execution of complex tasks,such as the MUSIC algorithm.This thesis,focusing on the super-resolution direction platform,mainly carries out the following work:The key technologies of super-resolution direction-finding system are studied,such as MUSIC algorithm technology,spatial smoothing principle,and amplitude-phase error correction principle etc,and use MATLB to do corresponding simulation and analysis.The platform structure and software and hardware design schemes of direction finding system are given.The implementation of the super-resolution real-time direction finding system is designed,and the hardware implementation framework is given.The hardware structure of FPGA(Field Programmable Gate Array)+ DSP(Digital Signal Process)are adopted.FPGA is used as signal preprocessing and system control,and DSP is the core of digital signal processing.The overall work flow of the super-resolution direction finding system and the software division of FPGA and DSP are given.Designed FPGA software functions and communication interface functions.The software functions include sampling,initialization,covariance calculation,etc.The communication interface module includes SPI digital control,protocol transmission of upper computer,and EMIF bus transmission.Designed the main software function modules of DSP,including software online upgrade,amplitude and phase correction and implementation of MUSIC algorithm,etc.The real-time performance of the super-resolution real-time direction finding system is studied.The reason why the system in this thesis can achieve real-time direction finding is because of the following advanced technologies: 1)The parallel pipeline data processing structure is used to calculate the covariance in the FPGA;2)The spectrum peak search is performed in collaboration with multiple cores.The realization result of the super-resolution real-time direction finding system is shown.Including the result of hardware circuit realization,host computer realization and MUSIC algorithm realization.And self-built indoor test platform to carry out relevant performance tests,compare and analyze the estimated results,and conclude that the system design meets the design requirements.