Research Of Adaptive RLS Algorithm And The Application Of The Noise Filtering In Periodic Signal

All of the systems are inevitably influenced by kinds of the noises. One of the hotresearch topic in recent years is how to effectively eliminate and reduce the noise. Noisesuppression method can be divided into two parts: passive noise suppression and active noisesuppression.Passive noise suppression method gets at the acoustic materials and structure ofacoustics, effectively suppresses the high frequency noise, but the effect of suppressing thelow frequency noise is not good, and not only the equipment is bulky,but also the volume ishuge,difficult to install. Active noise suppression method can solve the problem of the lowfrequency noise suppression. The system of the volume is small, and the weight is light. Withthe development of the digital signal processing technology and theory of the control system,the adaptive technology becomes the main research direction of active noise suppressiontechnique. Adaptive filter is a kind of developed optimal filtering method in the nearly thirtyyears.It developed basis on the wiener filtering, Kalman filter and some other linear filters. Ithas a stronger suitability and more optimal filtering performance.This article mainly introduces the basic principles of adaptive filter, researching the mostbasic two algorithms of the adaptive algorithm-LMS and RLS deeply, by analysis theinfluence to the algorithm by the characterristics of typical algorithms and each parameter andcomparing the advantages and disadvantages of typical algorithms. Adaptive LMS filter’sstructure is simple and easy to be realized, but its lack of point is the very slow convergencespeed and the poor anti-interference. Based on the least square criteria, RLS algorithmrecursive estimates and updates the Autocorrelation inverse Matrix of the input signal, whoseconvergence speed is faster and tracking ability is stronger. Its convergence and the spectrumcharacteristics of the input signal are independent. The system is more stable. By comparingthe advantages and disadvantages of various RLS filtering algorithm, we improve the existingRLS algorithm with forgetting factor and realize the auto-tuned filter. For validateing thefeasibility of the algorithm, we analysis the algorithm’s tracking ability、convergence and theanti-interference ability in theory. By comparative simulating them by the Matlab software,we has confirmed the superiority of the improvement RLS algorithm. In the end, we programthe improved algorithm by the VHDL in QuartusII software development platform, and useEP1C6Q240C8of the Cyclone series in the SOPC box to implement this algorithm. Eventually we improve the system and form a more mature system of periodic signalacquisition. It is applicable to filtering process of the weak signal and low frequency cycle.And it can be used independently.