Design And Application Of Radar Semantic Structure Analysis Algorithm

An important content of radar signal semantic structure analysis is that to identify a relatively specific meaning of high-level operation mode from low-level radar signal sequence. To mark a start point of some kind of operation mode in the signal sequence is a common method, which can only obtain the border of the operation mode, but not the internal structure information. Radar semantic structure analysis is to study the internal structure of the radar operation mode, which means that to study how the radar operation mode combine radar signal sequence form low-level to each level, then finally compose different operation mode. It not only provides another way for pattern recognition problems, but also provides useful information for analyzing the commonness and difference of different operation mode, and high-level semantic interpretation for radar signal. Therefore, radar semantic structure analysis has important theoretical and application value.Were reviewed in this paper on the basis of previous work, the radar semantic structure is divided into four levels: low level of radar and radar,radar, radar pulse sequence word phrase sentence (working mode),reference to human natural language some ideas of hierarchy analysis,carried out the following work:In this paper, on the basis of previous work, the radar semantic structure is divided into four levels: low level radar pulse sequence, radar words, radar phrases and radar sentences (operation mode). Reference to some ideas of hierarchical structure analysis for natural language, this paper carried out the work as follow:Based on the analysis of pulse sequence in different operation mode,this paper put forward a method for constructing radar words based on Hidden Markov Model (HMM), and designed the HMM radar words recognition algorithm in radar signal sequence, which cloud transform pulse sequence to concise radar words sequence. Experimental results show that, the radar words constructing and recognition algorithm in this paper has well performance in multiple operation modes using their own radar words, and multiple operation modes sharing the same set of radar words as well.This paper put forward a set of data driven Probabilistic Context Free Grammar (PCFG) rule-extraction method based on the radar words, which used the radar words bigram data to obtain the PCFG ruled that satisfy Chomsky paradigm. The extraction PCFG rules is a set that constituted from radar words to radar phrases and radar sentences. Based on that rule set, we can build the hierarchical structure tree of radar sentences(operation mode). The experimental results showed its effectiveness.Based on the aforementioned method, we created a radar signal structure analysis system, which includes semantic state analysis function,and helps to recognize radar words and extract PCFG rules as well.