Applications Of Fast Filter Bank In Real Time Spectrum Analyzer

Wireless communication technology developing rapidly at present.In the military field,Many power devices such as various radio,radar,aircraft avionics pod,was put in a relatively narrow space of the battlefield,It is easy to form the complex battlefield electromagnetic environment.Characterized by complicated composition of frequency and it's uncertain survival time.The traditional frequency sweep measurement is based on the way of frequency scanningusing.It is actually a kind of tuning superheterodyne receiver.Due to the structural defects,the frequency out of the current scanning spot can not be found,The transient signal may have been missed.Real-time spectrum analyzer can solve the above problem.The basic principle of frequency conversion is as follows:First the signal down frequency,then conversion to digital,and then converted to frequency components through the FFT algorithm.But FFT/IFFT exist two main disadvantages:first,with the continuous increase or decrease of sample points FFT/IFFT transformation,makes the bandwidth of the main lobe is smaller,but has no effect on the first side lobe attenuation is,to increase the side-lobe attenuation,usually before IFFT/FFT transform to add all kinds of window function,this operation needs to consume large amounts of resources.Secondly,FFT/IFFT filter group of side-lobe attenuation increases at the same time have become the main lobe fuzzy problem,which limits the further improve the performance of the FFT/IFFT.So you need a good filter performance group instead of FFT/IFFT,the filter to have larger side-lobe attenuation,demand has lower implementation complexity at the same time.FFB like FFT/IFFT,it was implemented through a set of parallel tree structure of filter bank,per child level filter of the prototype filter transfer function is the same,can take advantage of the prototype filter transfer function,through the way of factor substitution,we can get the transfer function.In order to improve the performance of FFT/IFFT,FFB use advanced prototype filter to replace the corresponding original prototype filter,which can greatly increase its side-lobe attenuation.This article mainly aimed at the real-time spectral analysis algorithm based on fast filter group is studied,the main work is as follows:First,the method about real-time spectrum detection of the mainstream is introduced.Analyzed the limitation of the FFT and the superiority of FFB.Secondly,the basic theory of FFB is studied.Analyzed from the Angle of algorithm level,we can see it has good frequency response characteristics and the advantages of lower complexity compared with FFT.Thirdly,discussed in detail the relation between the various sub filter of the FFB,The prototype filter optimization design method and the concrete implementation steps is given.Fourth,put forward a kind of improved type FFB based on the variable digital filter(VDF).This filter bank can be flexibly adjust the subchannel bandwidth and center frequency of the subchannel,at the same time also ensure that have enough frequency resolution and low system complexity.Finally,In this thesis proposes a frequency estimation algorithm based on FFB and adaptive notch filter.The implementation method is greatly to reduce the system response time compared with the existing frequency estimation algorithm on the basis of adaptive notch filter,but also guarantee with sufficient frequency resolution and low system complexity.