Weak Signal Detection And Separation In Radio Spectrum Monitoring

The radio spectrum management is becoming more important in today’s wirelesscommunication and the spectrum management is a core part in the whole work. Using the rarespectrum resource efficiently, optimizing ever field’s utilization rate of spectrum and avoidingwaste and illegal use of the spectrum has been imminent. In the radio spectrum monitoringmanagement, the existing of weak signal may not be ignored and it may bring more obstacles.In order to analysis the signal in spectrum area, we transform the received signals tofrequency spectrum with the software of LabWindows/CVI. According to the background ofradio spectrum monitoring management, we discuss the weak signal under noise and nearbybig signal.1. Aiming at the problem of detecting weak signal submerged by the noise, we presentsuperposition method in time domain and frequency domain according to the specialstatistical character of noise. In algorithm, we provide an improved iterative superpositionalgorithm and take first simulation and then transplant design ideas.2. For the question of weak signal existing nearly by big signals, if we transform themusing the traditional FFT, a blunt peak is easy to form and it is difficult to check weak signal.AR linear prediction model is being adopted to improve the resolving power of spectrum. Theselection of model’s order is a pivotal segment in the forecasting process. When weak signalis detected at one frequency point, we further study the multiple signal classification (MUSIC)algorithm using the advantage of DOA. Signal space vector is constructed and then beencharacteristic decomposition. Detecting of weak signals and wave direction at the samefrequency may avoid the omission of the number of weak signal.The Matlab and Labwindows/CVI simulation results show that the improved iterativesuperposition algorithm is good and can detect weak signal easy. When AR and MUSIC areused in combination, they are good for improving the spectral resolution checking thenumbers of signal. However, optimal order selection is influenced by the number of signal,noise amplitude and other multiple factors.Transplanting the algorithm to hardware environment, the actual algorithm validationshows that the improved iterative superposition algorithm may obtain good results whendetecting weak signal under noise and the algorithm is easy to achieve. The FrequencyDomain Superposition is better. AR model is used to verify the actual receive data at last.When its order is suitable, it can greatly improve the spectral resolution and separate the weaksignal from the strength ones.