Research On Frequency Estimation Algorithm Of Sinusoid Signal And Hardware System Implementation

Frequency estimation has a common application in practical engineering,and it has always been a research focus in the field of digital signal processing.The frequency estimation of sinusoidal signals in the context of Gaussian white noise is a classical problem.The estimation accuracy,the estimation range and the computational complexity are the main indicators to weigh the performance of a frequency estimation algorithm.The sine signal frequency estimation algorithm can be mainly divided into two major categories:time domain algorithm and frequency domain algorithm.The DFT-based frequency domain algorithm has attracted the attention of researchers because of its remarkable feature of low computational complexity.In this paper,the following studies have been conducted for frequency estimation of different forms of signals:For the frequency estimation of complex sinusoidal signals,two new estimation algorithms are proposed in this paper.The first one is to propose an algorithm based on N-point sampling signal interpolation DFT.According to the formula,two frequency estimation methods with different range of application are deduced,and we choose the estimation method to obtain the result of frequency estimation by the threshold.Compared with the existing N sampling points methods,the proposed frequency estimation algorithm has higher accuracy,and has lower computational complexity than other complex algorithms,which is a trade-off between accuracy and computational complexity.The second method proposed is zero-padded 2N point frequency estimation algorithm.The algorithm has higher estimation accuracy,and it can further improve the estimation accuracy of the algorithm by iterative calculation.A frequency estimation algorithm for real sinusoidal signals is studied.Because the real sinusoidal signal and complex sinusoidal signal have different frequency domain characteristics,the algorithms for frequency estimation of these two signals are different and not universal.In this regard,we use an independent estimation iterative subtraction strategy to get the estimation result.Through iterative calculation,we estimate the frequency and amplitude of the signal respectively.We subtract the negative frequency part of the interference term in the calculation process and combine the frequency estimation of the complex sine signal to get the result.The simulation results show that the algorithm has accurate estimation and stable performance.This paper also carries out frequency estimation for special signals.For sinusoidal signals containing second harmonics,we adopts an iterative de-interference strategy,and estimates the frequency and amplitude of signals separately during the iteration process.The influence of the interference term of the harmonic frequency is subtracted,the estimation result is constantly modified,and the estimation result of the main frequency is finally obtained.The frequency estimation is performed for orthogonal signals whose amplitudes are different and whose phases are not strictly orthogonal.After calculation,the frequency estimation of these two signals can be combined with the idea of an estimation algorithm for real sine signals and harmonic sinusoidal signals,and iteratively complete the accurate estimation of the frequency.The simulation results show that the algorithm has a higher estimation accuracy and a stable estimation effect.FPGA implementation of frequency estimation algorithm for orthogonal sinusoidal signals is performed.This paper completes the overall process design of the algorithm,including the design of functional modules and the overall simulation,the selection and design of digital circuit devices,and the implementation of the algorithm in the FPGA is completed.The experimental results show that the algorithm has a good estimation effect on the frequency estimation of orthogonal sinusoidal signals.The experimental results of the algorithm are basically consistent with the simulation results,and are slightly different when the SNR is high.Finally,the error between the experimental results and simulation results and the cause of the error are analyzed in the paper.