| In recent years, with the rapid development of radio career, radio equipment is widelyused in various industries and fields, and it promote the development of our social economygreatly. However, some people using the rapid speed of radio communication for illegalactivities. Among them using radio equipment for cheating is especially, and thismisdemeanor damages the fairness and confidentiality of the exam severely. Because thecheating signal has such features as widespread of frequency band, unset of frequency point,randomness of the signal power and occurrence time, and the short duration of signal. So it isa challenge faced in the radio monitoring and management to find the cheating signalefficiently, quickly and accurately.In this paper, in the field of monitoring of cheating signal, we use the fuzzy patternrecognition method to classify and identify the radio signal. Using the improved ISODATAclustering algorithm (FCM algorithm), and going ahead for fuzzy clustering analysisaccording to the characteristics of cheating signal. The main research results of this paper canbe summarized as follows:1. Preprocessing the spectrum data of radio signal, and giving out some datacharacteristics for describing the cheating signal, setting up the signal feature database.2. Let the fuzzy pattern recognition method apply to the recognition of cheat signal, andput forward a new improved ISODATA algorithm based on weighting. We use the newalgorithm to calculate the fuzzy classification matrix and the clustering center matrix of thesignals.3. The weighted absolute distance is used to calculate the distance between the unknownsignal and each row vector of the best cluster center matrix. According to the minimumprinciple of distance, the type of the signal can be determined.4. Let the suspicious signal data which is intercepted in the national graduate entranceexam of xihua university in2012as the experimental data. Experiment verifies thecorrectness and feasibility of the theory. |
PDF Full Text Request