Research on radio signal feature extraction and classification method based on principal component analysisAuthor(s): Liu Yang
Due to the fast development of technology, especially the rapid changes in radio technology, peopleÂs lives are increasingly depending on radio technology. Radio spectrum enriches people's lives, which makes spectrum resources particularly precious. However, at present, there is a lot of abuse of the spectrum in people's lives, which not only leads to a substantial waste of spectrum resources, also sometimes poses a threat to human health. Further, this will cause disruptions to the whole communications industry and thus restrict the rapid socio-economic development. In order to solve this thorny problem, well-known experts and scholars began to study radio signal feature extraction and classification methods. In recent years, many scholars have been trying to settle the issue (namely abnormal signal analysis) in the area of communication by using fuzzy set theory and neural networkï¼and made great achievements. This paper adopts principal component analysis method to extract the features of radio signals, and improves the previous extraction methods by combining entropy method with fuzzy c-means, and applies weighting method to label the importance of each feature in order to solve problem of uneven contribution in classification. It is proved that the improved method is of high efficiency, high speed, capable of determining categories of abnormal signals accurately and efficiently so as to guarantee the integrity and reliability of the information in the communication link, with great practical value.