Design And Implementation Of Embedded Real-time Processing Software For Aircraft Early Warning Radar

This thesis is based on a aircraft warning radar project,focusing on the theoretical re-search and program design of on-board multi-target tracking algorithms and real-time data processing software,based on the NVIDIA TX2 embedded hardware platform and Linux operating system.A set of doppler-assisted multi-array radar data processing software is developed.The core algorithm of the software includes measurement data pre-processing,multi-radar track initiation,multi-target data association and track termination algorithms.The main function is a radar data processing platform with multi-target tracking and track management capabilities,integrating multiple functions such as network communication and radar control.The main work of this thesis includes:Firstly,the modeling of the on-board radar system is studied,and the theoretical algo-rithms and engineering implementation of distance ambiguity,DBSCAN clustering algo-rithm,coordinate transformation and system error configuration are analyzed.Then based on single radar and multiple radars tracking environment,sequential processing track ini-tiation algorithm represented by logical method and the batch processing track initiation algorithm represented by Hough transform are studied.The impact of multiple radars on track initiation is analyzed through simulation experiments.Based on this,the doppler velocity gate is derived,and a doppler velocity-assisted multi-radar logical method track initiation is proposed.Based on the application scenario of aircraft warning radar in this article,the influence of multi-Doppler velocity gate on the selection of logical method slid-ing window size is analyzed.Finally,a doppler velocity-assisted multi-radar cheap JPDA algorithm is derived by introducing this multi-Doppler velocity gate into multi-target data association.Moreover,the important position and tasks of the data processing software in the aircraft warning radar system are analyzed,and a multi-array on-board radar data pro-cessing framework with object-oriented characteristics is designed based on the tasks.It integrated various functions such as measurement data pre-processing,track initiation,multi-target data association,tracking filtering,track termination,and track output.Based on the object-oriented design concept,these functional modules are implemented sequen-tially from partial to whole,and the framework achieved high scalability and low coupling through the way of abstract class and implementation class.And the radar control function is implemented,and a state machine is designed to complete functions such as radar state switching,parameter configuration,and real-time state feedback.By building a network communication and multi-threading framework,the radar main control function and radar data processing function are integrated,while ensuring memory safety and software running efficiency.Finally,a test platform is built,testing the various functions of the software with simulation data,and then evaluating the validity of the proposed algorithms and data pro-cessing software through field experiments and functional indicators testing.