Estimation of Doppler Spectrum Parameters - CRAHI估计多普勒频谱参数crahi.pptVIP

Estimation of Doppler Spectrum Parameters Comparison between FFT-based processing and Adaptive Filtering Processing Contents Signal and noise models of the adaptive processing Adaptive filter estimation Implementation of the two algorithms Comparison of the two algorithms Model of the atmospheric radar signal For each range cell: Complex autoregressive series of first order with unknown coefficient Spectrum symmetric relative to average Doppler frequency Calculation of Doppler spectrum parameters Doppler spectrum parameters Mean Doppler frequency: Doppler spectrum width: Simulated Signal Simulated signal: Noise: uncorrelated additive normal complex white noise Adaptive filter estimation Objective: Suppress noise Accurately estimate a Adaptive filter estimation Implementation of the algorithms Comparison: Processing speed Number of operations: FFT: 6N log2 N Adaptive: 41 N Advantage in real time operations ! Convergence of spectral moments Overestimation of Doppler spectrum width Estimation of weak signals Comparison: Real Data Conclusion Adaptive Doppler processing: Good estimation of the Reflectivity and mean Doppler velocity Overestimation of Doppler spectrum width Faster signal processing than FFT processing Estimation of weaker signals Simple signal model: Bi-modal atmospheric signal? ERAD 2006: 18-22 Sept. 2006 * J. Figueras i Ventura1, M. Pinsky2, A. Sterking3, A. Khain2, H.W.J. Russchenberg1 1 IRCTR-TU Delft 2 Institute of Earth Sciences-The Hebrew University of Jerusalem 3 Weizmann Institute of Science Recurrent estimation of a Mean Frequency: 0.5 Doppler width: 0.4 NaN 0.45 0.3 Doppler width NaN 1.03 1 Mean Velocity NaN 2.85 2.75 Signal Power FFT 6 dB Clipping Adaptive Given Parameter SNR=0.09 Clipping: ERAD 2006: 18-22 Sept. 2006 *