Application Of Unsupervised Clustering Algorithm To Emitter Signals Analysis

Recognition of radar emitter signals is an important symbol to measure the technical level of radar countermeasure in modern electronic countermeasure, besides, it is an urgent problem to solve. With the increasing development of military technology and more complex of electromagnetic environment, some traditional classification methods have been unable to meet the requirements of recognizing unknown radar emitter signals satisfactorily.Unsupervised learning is a good method to solve the problem of recognition of unknown radar emitter signal. Support vector clustering (SVC) is a kind of unsupervised clustering method which based on support vector machine. SVC can obtain cluster boundaries of arbitrary shape without a priori knowledge of the data sets. Kernel parameter of the SVC algorithm plays an important role in clustering formation, which affects the boundary and shape of cluster. How to find the optimal kernel parameter value is the key to identify emitter signal accurately. SVC not only has a high time complexity but also gets low recognition rate when distribution of the data set is complex and uneven. So we do some research on how to get optimal parameter value, at the meanwhile improve the recognition rate with a lower time complexity when distribution of the data set is complex and uneven as to get fast and accurate recognition of radar emitter signal. The main work and research results of this paper as follows:1. We discuss several unsupervised clustering algorithms, emphatically introducing K-means and Fuzzry C-means (FCM).Then we compare and analyze their classification results for data set, iris and glass.2. SVC is applied in radar emitter signal recognition. The experimental results show that SVC not only has a high time complexity but also gets low recognition rate when distribution of the data set is complex and uneven.3. We propose a novel validity measures QEA-SVC algorithm to find optimal SVC kernel parameter value.We use the automatic optimization function of Quantum-Inspried Evolutionary Algotithm (QEA) to find the SVC kernel parameter value, which makes data set in an optimal clustering formation. QEA-SVC algorithm is applied in data set, iris and radar emitter signal recognition. The experimental results show that QEA-SVC algorithm can find SVC kernel parameter value effictively.4. Combining the advantages of SVC and FCM, a novel SVC based on FCM (FC-SVC) is presented. First, we use FCM to divide the data set into several subsets. Then, SVC is employed to cluster each subset. FC-SVC can decompose a large quadratic-programming problem into smaller ones. So, it can reduce the computational cost. When the data set is divided into several subsets, more suitable parameters can be easily found for each subset. The introduced method is tested by using data set and radar emitter signals. The experimental results show that FC-SVC can obtain better clustering performances in terms of computational efforts and correct recognition rates than other methods.