The Study On Algorithms And Applications Of Noise Cancelling Technology Based On Adaptive Filter

This paper firstly studies the basic principles, system structure and practical applications of the adaptive filter. It analysis the Least Mean Square(LMS)algorithm which is based on the Minimum Mean Square Error and Recursive Least Squares(RLS) algorithm which makes the Least Squares as criteria. Then analysis the advantages and disadvantages of the two algorithms and two algorithms are compared. A new algorithm of variable setep size LMS was presented, aiming at solving the disadvantage of the traditional algorithms’ convergence speed contradicting its steady-state error. On the other hand, this algorithm solves the questions of system get relatively large fluctuations when the error of system close to zero and system performance degradated due to low SNR. The simulation results show the performance which exceeds other algorithms.Secondly, the paper studies the theory of adaptive noise cancelling system which based on adaptive algorithm. The essence of adaptive noise cancelling system is eliminating noise effectively by adaptive filter theory. Then introduce the whole system by analysising structure, properties and metrics. The performance of traditional algorithms will degrade dramatically when system in environment with impulsive noise, at the same time, system often obtains the channel signal difficultly. This paper proposes a new algorithm to solve this two questions, farther more, analyses and compares with the traditional algorithm. The computer simulation results show the new algorithm owns better de-noising effect than traditional algorithm in extracting periodic signal. It also overcomes the deterioration problems of traditional algorithm when system with impulse noise.Finally, on the basis of studying above theoretical and simulation, the paper employe a high speed processing unit which type number is TMS320VC5509 A, an audio decoder chip TLV320AIC23 which own high performance and other related peripheral modules to build an adaptive noise cancellation system. Realize the traditional LMS algorithm, RLS algorithm and the new algorithm by integrated system development environment CCS4.2 which choosed in the simulators mode. Then the system chooses hardware simulators mode, uses the adaptive LMS algorithm which simple to hardware realization to eliminate white Gaussian noise and factory noise to get pure audio signal, all noise are obtained from noise92 libiary. Finally, realizing real-time processing by the hardware simulators download link based on designed noise cancellations system. The paper introduces the whole system components in detail, show the connection between the DSP module and the AIC23 module, the software flow chart and the test result. All of these have good reference value.