Distributed Radar Parallel Processing Implementation And Field Experiment Research

Distributed radar has many advantages,such as high flexibility,strong detection capabilities,wide detection range,strong anti-jamming and anti-interception capabilities,etc.,so it has become a research hotspot in the current radar field.Aiming at the key technical problems faced by the wide-area distributed radar fusion detection system in actual engineering implementation,this thesis conducts application research from system model,to processing algorithm,to real-time processing,and finally to field test verification.The specific contributions of this thesis are as follows:1.The wide-area distributed radar signal model and signal strategy are studied.For the wide-area ultra-sparse radar node system,conducted modeling of received signal echoes based on near-field models and studied the signal processing process of distributed radar systems.In particular,a near-field comparison-amplitude goniometric algorithm for wide-area ultra-sparse systems is proposed.Simulation experiments show that the angle measurement algorithm is still effective in near-field scenarios.A distributed radar realtime processing strategy based on guidance information was proposed,and its feasibility was verified through simulation.2.The calibration method of amplitude and phase error of ultra-sparse node system is studied.For wide-area ultra-sparse radar node systems,the influence of amplitude and phase errors between nodes on coherent synthesis is analyzed.Based on the distributed radar near-field signal model,a method for amplitude and phase error correction based on auxiliary array elements is studied,and a distributed radar amplitude and phase error self-correction method that is easy to implement in engineering is also studied.The effectiveness of this method is verified through field experiments.3.The realization of real-time processing based on multi-server multi-GPU(Graphics Processing Unit,GPU)platform is studied.Two signal processing task partitioning schemes were proposed and their advantages and disadvantages were compared.Specifically,the implementation process of a node parallel scheme on a multiserver and multi GPU platform was studied,and the signal processing implementation within a single server and synchronization processing between multiple servers were provided.The real-time performance of the system was verified by simulating data sources.4.The field test verification of the wide-area distributed radar system is carried out.Use the wide-area distributed radar implemented in this thesis to detect civil aviation aircraft in real time.The detection results show that on the premise that a single node does not have the ability to detect targets alone,multiple nodes can achieve real-time tracking and detection of multiple targets through joint coherent processing,which verifies the feasibility of the entire system implemented in this thesis.