Research And Application Of Data Cleansing In Multi-radar Data Fusion Algorithm

With the increasing air traffic flow,the volume of reported data from the radar network under the surveillance of the air traffic control system is a large-scale increase,which has particularly highlighted the quality problems in radar data.It is vital for the air traffic control information processing systems to improve the quality of radar data.Data cleansing technology is widely used to improve data quality.Among them,missing value cleansing and,detection and processing of approximately duplicated records are indispensable research contents in application.This paper firstly studies and implements the subflows of the multi-radar information fusion process,which are spatial alignment,UTM projection,time alignment and data fusion.The fast training function is designed and implemented in the data fusion algorithm.Secondly,based on data cleansing technology,the data cleansing sub-process is studied and applied to the original information fusion process.The research contents are as follows:according to the attribute characteristics of radar data,this paper proposes a missing value cleansing method of radar data based on regression interpolation;the improvement of dynamic stratification of hierarchical mean interpolation method is implemented,and the dynamic layered mean interpolation method for radar data is proposed;the narrowly defined repeated records in the radar data are cleansed using the latest processing principles,and a cleaning method that combines integrity priority,comprehensive evaluation and the latest processing principles for different records in a narrow sense is proposed.Finally,based on the actual aeronautical data,the design and control experiments were carried out to study the application effect of the cleansing method of radar data in the multi-radar data fusion algorithm,and the radar data quality assessment method was used to evaluate the quality of the experimental data before and after the data cleansing.The data quality assessment results show that the radar data cleansing method can improve the radar data quality from the attribute dimension of data integrity,flight target elevation accuracy and repeatability attribute dimensions.The experimental results are analyzed and the following conclusions are drawn:the application of dynamic layered mean interpolation method of radar data can significantly improve the training speed and training efficiency of multi-radar data fusion algorithm;the application of the radar data missing value cleansing method based on regression interpolation can improve the multi-radar data fusion algorithm from the three aspects of the accuracy and effectiveness of the fusion performance,training speed and training efficiency.Therefore,the application of the radar data cleansing methods proposed by this paper in the multi-radar data fusion algorithm is helpful to improve the performance of the fusion algorithm,and the radar data cleansing methods have certain feasibility in the field of radar information processing.