초록 ▼

Systems and methods that adapt to the weather and clutter in a weather radar signal and apply a frequency domain approach that uses a Gaussian clutter model to remove ground clutter over a variable number of spectral components that is dependent on the assumed clutter width, signal power, Nyquist in

Systems and methods that adapt to the weather and clutter in a weather radar signal and apply a frequency domain approach that uses a Gaussian clutter model to remove ground clutter over a variable number of spectral components that is dependent on the assumed clutter width, signal power, Nyquist interval and number of samples. A Gaussian weather model is used to iteratively interpolate over the components that have been removed, if any, thus restoring any overlapped weather spectrum with minimal bias caused by the clutter filter. The system uses a DFT approach. In one embodiment, the process is first performed with a Hamming window and then, based on the outcome, the Hamming results are kept or a portion of the process is repeated with a different window. Thus, proper windows are utilized to minimize the negative impact of more aggressive windows.

대표청구항 ▼

What is claimed is: 1. A system for processing a radar signal comprising: (a) a memory storing software and echo data related to a transmitted pulse, wherein the echo data includes an input time series; and (b) a processor for running the software, the processor being in communication with the memo

What is claimed is: 1. A system for processing a radar signal comprising: (a) a memory storing software and echo data related to a transmitted pulse, wherein the echo data includes an input time series; and (b) a processor for running the software, the processor being in communication with the memory, wherein the processor is operative to: (i) apply a window to the input time series; (ii) apply a discrete Fourier transform to the windowed input time series to obtain a Doppler power spectrum; (iii) filter the Doppler power spectrum to remove clutter points; and (iv) perform an iterative interpolation using a Gaussian model to replace the artifacts that are removed by filtering the Doppler power spectrum. 2. A system as recited in claim 1, wherein the processor is further operative to select and apply a window based upon clutter in the echo data. 3. A system as recited in claim 1, wherein the processor is further operative to set a gain of the window to preserve total power. 4. A system as recited in claim 1, wherein the processor is further operative to output moments of the Doppler power spectrum to an application processor, which generates a display for meteorological interpretation. 5. A system as recited in claim 4, wherein the moments are intensity, mean radial velocity and spectrum width of weather targets in a beam for each range increment. 6. A system as recited in claim 1, wherein the processor is further operative to apply a FFT approach when a number of samples is a power of two. 7. A system as recited in claim 1, wherein the window is an aggressive low side-lobe window. 8. A system as recited in claim 7, wherein the processor is further operative to automatically adjust a width of a clutter filter based on radar wavelength, pulse repetition frequency, number of pulse samples and strength of the clutter. 9. A system as recited in claim 1, wherein the processor is further operative to calculate an autocorrelation at lag zero without clutter filtering, the autocorrelation being computed by taking an inverse DFT of the Doppler power spectrum in a frequency domain. 10. A system as recited in claim 1, wherein the artifacts are selected from the group consisting of ground clutter, aircraft, interference from other radars, and known physical structures. 11. A method for processing data generated by a radar system comprising the steps of: (a) applying a first guess weighting function to input time series data and then applying a discrete Fourier transform (D FT) to the windowed input time series data to generate a Doppler power spectrum; (b) determining and removing an optimal number of clutter points based on a Gaussian model of clutter, an assumed clutter width and a measured receiver noise power; (c) calculating moments and using the moments to fit a Gaussian function to remaining weather points to make a next guess Doppler spectrum; (d) recalculating the moments based on the next guess Doppler spectrum; (e) determining if the recalculated moments obtained in step (d) based on the next guess Doppler power spectrum are acceptably close to the moments calculated by in step (c) by comparison of differences between the moments to a user-defined tolerance; (f) if the differences between the moments of the Doppler power spectrum are not acceptable, recalculating the moments of the Doppler power spectrum based on moment estimates and repeating steps (c), (d) and (e) until the next guess and previous guess of the moments agree to within the user-defined tolerance; and (g) checking for usage of an appropriate window and noise power value and recalculating the moments by repeating steps (a)-(f), wherein the appropriate window is determined according to the following: i) if a clutter-to-signal ratio (CSR)>a first adjustable parameter, the first adjustable parameter typically being about 40 dB, using a Blackman window and dynamic noise calculation and using results thereof for the moments of step (d); and ii) if the CSR>a second adjustable parameter, the second adjustable parameter typically being about 20 dB, using a Blackman window, then if CSR>25 dB using the Blackman results for the moments of step (d); and iii) if the CSR<a third adjustable parameter, the third adjustable parameter typically being about 2.5 dB, using a rectangular window, then if the CSR<a fourth adjustable parameter, the fourth adjustable parameter typically being about 1 dB, using the rectangular window; otherwise if i), ii) and iii) do not occur, using the results of the first guess weighting function. 12. A method as recited in claim 11, wherein the first guess weighting function is a Hamming window. 13. A method as recited in claim 11, further comprising the step of performing an inverse DFT to obtain an autocorrelation at lags 0, 1 and applying a Gaussian model using the calculated moments to determine a substitution value for each of the clutter points that were removed. 14. A method as recited in claim 13, wherein the clutter points are removed by fitting a Gaussian curve to three central spectrum points of the Doppler power spectrum and discarding points within an intersection of the Gaussian curve and a noise level. 15. A method as recited in claim 14, wherein the Gaussian curve has a nominal spectrum based on the number of spectrum samples, PRF and wavelength. 16. A method as recited in claim 11, wherein removal of the clutter points is based on the Doppler power spectrum, assumed clutter width and the noise level, and the noise level is known from noise measurement. 17. A method as recited in claim 11, wherein if a sum of three central components is less than a noise power for the Doppler power spectrum, then all of the moments are calculated using a rectangular window and if the sum is only slightly larger than the noise level, then only a most central component is removed. 18. A method as recited in claim 17, when the clutter power is very strong, further comprising the step of computing the noise power using a dynamic approach of sorting the Doppler spectrum components in order of power to define two regions: a noise region; and a signal/clutter region, wherein the noise level and a transition between the two regions is determined by first summing a power in a range of 5% to 40%, the power is used to determine a boundary noise level by comparing with the power corresponding to a theoretical curve, then, the power is summed beyond the 40% point for both an actual and a theoretical rank spectra, at a point where the actual power sum exceeds the theoretical value by 2 dB determines a boundary between the noise region and the signal/clutter region to generate two outputs: a spectrum noise level; and a list of components that are either signal or clutter. 19. A processor for a weather radar system comprising: memory for storing data and instructional code; a processor in communication with the memory for accessing the data and executing the instructional code a software module stored in the memory and including instructions for the processor to execute the following steps of: i) applying a window weighting function to I and Q values to remove ground clutter prior to generating a Doppler power spectrum; ii) generating the Doppler power spectrum; iii) performing a discrete Fourier transform on the Doppler power spectrum; iv) determining a noise power in order to determine clutter points; v) removing the clutter points; and vi) replacing the clutter points with signal components to create the resulting Doppler power spectrum. 20. A processor as recited in claim 19, wherein the module is further operative to iterate through steps iv) through vi). 21. A system as recited in claim 1, wherein the window is a rectangular window. 22. A method as recited in claim 11, wherein the user-defined tolerance is less than 0.2 dB.