The Improved Frequency Estimation Algorithm Using DFT Interpolation Algorithm And Its Implementation On DSP

Frequency estimation is an important research topic in the digital signal processing(DSP)field,has already been widely used in military,communications,speech processing and recognition,biological medicine,power systems,and many other fields.And get more and more attention from experts and scholars.So the study of frequency estimation algorithm has great significance and practical value.After years of development,experts and scholars have put forward a variety of frequency estimation algorithms,including maximum likelihood ratio estimation,modern space decomposition frequency estimation and all kinds of spectrum correction method,etc.Maximum likelihood estimation of sinusoidal signal frequency parameters can be obtained from measured signal cycle diagram maximum extremum,but this method is time-consuming.Modern frequency estimation based on space decomposition has very high estimation precision,but its implementation process is complex.Due to the discrete Fourier transform(DFT)can be obtained by fast Fourier transform algorithm(FFT)rapidly,so it is very convenient for real-time processing field,and can apply the appropriate frequency spectrum interpolation correction technology to get high precision of frequency estimation results.This paper will mainly discuss and study the principle,implementation and application of the DFT frequency estimation algorithm.The main content includes the following several aspects:(1)Carryed out the theoretical research and simulation analysis of the Rife algorithm,Quinn algorithm and A&M iterative algorithm such classic DFT interpolation frequency estimation algorithm.(2)On the basis of these classical algorithms,an improved DFT interpolation frequency estimation algorithm is proposed,this algorithm first estimate a frequency error through Quinn algorithm as the error initial value of A&M iterative estimation algorithm,and then use the A&M algorithm to estimate a more accurate frequency result.The performance of this improved algorithm and several classical algorithms are compared,and the results show that this improved algorithm can effectively reduce the iterative number of A&M algorithm,which makes this improved algorithm has the characteristic of high efficiency with high precision.(3)Put forward the method to implement this improved algorithm on the DSP processor,and in order to improve the efficiency of the algorithm on the processor,this paper also puts forward an optimization method of A&M algorithm.(4)Applied this improved algorithm to the pulse detection system.Introduced various parts of the system,designed the hardware and software of the system,did the final testing of the system.To sum up,this paper first studied the frequency estimation algorithm,and then proposed an improved algorithm of high accuracy and high efficiency,and its implementation method is given on DSP processor,at last applied this frequency estimation algorithm to the pulse detection system in the ECG(electrocardiogram)monitoring field,did the design and testing of this system.