Open Radio Detection System Key Technology Research

In the limited space, there is limited radio spectrum resources, with the development of the social electronic technology, the number of the electronic equipment have dramatically increased, it made the electromagnetic environment deteriorated, in the modern time while the electronic equipment increasing tend to miniaturization and miniaturization, it came to the resulting that any other devices can be easily interference by other radio wave, people’s daily lives could be caused by these phenomena inconvenience, the worse is it could caused a security risk. For it can influence other devices, and cause dangerous accidents, we should manage it in a standard way. Radio wave of devices radiated detection as an important part of the compatibility testing in the electrical measuring, now commonly used the method of an anechoic chamber test method, the low detection efficiency of the anechoic chamber and high testing costs will become the obstacle of the production in the development of the electronics industry.This thesis is focused on the deficiencies of traditional methods, work in an open environment of the RF test methods based on hypothesis testing. The background signal samples and samples containing background signal and the signal under test, respectively, frame it in the time domain, reorganized each frame signal on the same frequency in the frequency domain. By the final we test the frequency two points off the shutdown sample set of statistical to determine the difference between the two samples collection, and then judge whether the two samples from the different distribution, and it lends to test equipment of the RF band. If the two samples sets from the same distribution, that there is no equipment under test’s frequency, on the contrary, that equipment under test’s frequency is detected. Mainly focus on the following arguments, a major problem in the open environment of the RF detection, is the background noise could interference the detection, affect the final test results. The fundamental deficiencies of the traditional frequency measurement are the low signal to noise ratio and the complex time-varying characteristics of the background noise, the two spectrum contrast to the environment is not accurate, and the uncertainties of the test can not be quantified. We assume that hypothesis testing as the core algorithm of the detection system to compensate for this deficiency. It used sample data to obtain sample statistics, and then use it to assume that the overall parameters of these data are reliable to a certain extent, and made to admit or reject the hypothetical judgment. The size of collected samples in the background RF signal mainly determines the cost of detection, simulation by testing different test methods such as parametric tests and nonparametric tests, and analysis of the different environments balance the test cost and test efficiency, to adjust the test under different background noise distribution. Finally, for complex variable background noise, the adaptive threshold is proposed to improve the constant threshold test method of hypothesis testing to detect the RF signal in low SNR. This method is based on data characteristics of the background signal and the signal to be detected with background sampling and signal frame’s spectrum analysis. T-test based on the normal population mean and unknown to the overall distribution of the KS-test results by the frequency of background signal spectrum extracted from the background signal threshold’s limit. Limit will be referred to remove the threshold by the detection signal and the background signal, using the same test methods to conduct statistical tests, comparative analysis of reference significantly with threshold to determine the RF the band of the devices under test.Based on parametric and nonparametric hypothesis testing methods it simulated in different signal to noise ratio of the signal detection, analyzed the test results show that the test in the normal background noise、parametric tests do a better job than the non-parametric tests; under the small sample size of the abnormal background noise, non-parametric test was significantly better than the parametric test; adaptive threshold achieves a excellent work compared to constant threshold.