Quinn algorithm is a low computational complexity and steady method which is widely used in frequency estimation of single real sinusoid in additive white Gaussian noise is given, But it has a problem of large variance of frequency estimation when the signal frequency is closed to the DFT ( Discrete Fourier Transform) discrete frequency in low SNR ( Signal Noise Ratio) ; However,compared to Quinn algorithm,the window-length changing phase difference correction algorithm has good frequency estimation precision in this case. But it also has a problem of large variance when the signal frequency is closed to the midpoint areas of two neighboring DFT discrete frequencies; Aimed at the characteristic, this paper proposes an conjoint algorithm which is a Quinn algorithm as weU as a window-length changing phase difference correction algorithm. The theoretical calculation process and related error formula are given out.From the result of computer Monte-Carlo simulation experiment and calculation amount analysis. We can see that the new algorithm has higher accuracy and stability in the frequency estimalion as well as low SNR threshold in the setting frequency range without increasing the calculation amount obviously. Its RMSE ( Root Mean Square Error) is close to CRLB( Cramer-Rao Lower Bound) in the whole frequency range. The performance of the new algorithm is better than Quinn algorithm and window-length changing phase difference correction algorithm, So it has some practical value for engineering.%Quinn算法是正弦波频率估计中应用广泛、计算量小且稳定性较好的算法,但是在低信噪比时当信号频率靠近离散傅立叶变换(DFT)的量化频率时,Quinn算法估计误差较大;而改变窗长相位差法在这种情况下具有较高的估计精度,但当信号频率位于两相邻离散傅立叶变换(DFT)的量化频率中心区域时,其估计误差很大.根据这两者特点,本文提出了一种基于Quinn算法和改进的改变窗长相位差法的正弦波频率估计综合算法,给出理论计算过程及相关误差公式.计算机蒙特卡罗模拟仿真实验和性能分析表明了本文算法在计算量增加不大的情况下,在设定频率范围内能够提高频率估计的精确度和稳定性,其均方误差接近克拉美罗限且具有较低信噪比门限,整体估计性能优于Quinn算法和改变窗长相位差法,具有工程实用价值.
展开▼
