Multiple Radar Time Series Similarity Matching Algorithm

The paper provide a new algorithm for multiple radar time series similarity matcing. The main difficulty for multiple radar time series similarity matching lies in huge data amount and too many factors will be considered. So, the previous algorithm on this area has existance of false dismissals and false alarms. The traditional methods mostly based on Euclidean distance and pay little attention to the influence of amplitude change, frequency persistence and internal structure of sequence. The matching method held out in the paper firstly devided the radar spatial data into certatin segments based on its´ spatial position. Then, the mutidemession time series in each data segments can maped to a time interval with certain length. The step can transform multiple radar time series to three-demession time series and greatly reduce the complexity of algorithm. In order to improve the accuracy of the method, some weighted factor with consideration of amplitude change, frequency persistence and internal structure of radar sequence is carried out to define an accuracy distance measure for two sequences. With the predefined distance measure, a multiple radar time series similarity measure modal and an algrothm for similarity matching is discussed in detail. At the end of the paper, the method in the paper is applied in actually radar observed data and proved its better performance.