Communications: directive radio wave systems and devices (e.g. – Radar for meteorological use
Reexamination Certificate
1999-08-13
2001-10-23
Sotomayor, John B. (Department: 3662)
Communications: directive radio wave systems and devices (e.g.,
Radar for meteorological use
C342S192000, C342S196000, C702S003000
Reexamination Certificate
active
06307500
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to Doppler measurement devices which are able to compute Doppler spectra as a function of range or time and, in particular, to a computation system which comprises an improved method of moment estimation, as applied to an automated meteorological monitoring system as the preferred embodiment of the invention, for the accurate real time detection of meteorological phenomena, such as winds, wind shear and turbulence.
PROBLEM
It is a problem in the field of spectra signal processing to extract valid data from the received signals. The signals can be radar, optical, infrared, tomography sonar in nature and the field of use can be meteorological monitoring, medical imaging, satellite imaging, automated mining, for example. The valid data represents a particular “signature” which must be identified and measured. The spectra received from the sensor(s) is corrupted by noise and the signature contained in the spectra is difficult to extract with present signal processing techniques.
It is particularly a problem in the field of automated meteorological monitoring systems to obtain timely, accurate and rapid estimates of meteorological phenomena that are extant in the region of space covered by the meteorological monitoring system. These monitoring systems typically use Doppler radar signals to ascertain the presence and locus of winds, wind shear and turbulence. To obtain timely, accurate and rapid estimates of these meteorological phenomena, the monitoring system must maximize the accuracy of the Doppler moments as well as the update rate and accuracy of the wind estimates. Existing automated meteorological monitoring systems are sometimes unable to achieve the desired results due to their inability to attain the accuracy required for these parameters.
Radar systems perform two primary functions: surveillance and tracking, and two secondary functions: target recognition and target measurement. For example, weather radar systems W as shown in
FIG. 2
are of the surveillance type and function to search a volume of space V to detect the presence and identify the locus of meteorological events M which can effect local activities, such as the flight operations at an airport. Modem weather radar systems W are pulsed Doppler radar systems that transmit a stream of fixed duration pulses P of radio frequency energy at repeated intervals, termed the pulse repetition period (PRP). The transmitted pulses P are reflected by a plurality of scatterers in the pulse volume M and the received return echo signals RE are processed to extract Doppler information.
Weather surveillance radars W continually scan a volume of space V. The antenna beamwidth, antenna scan rate and pulse repetition frequency of the radar transmitter determine the number of pulses transmitted per unit of time and hence the number of return echo signals received by the radar W. A typical weather surveillance radar W transmits a plurality of pulses during the time it takes the antenna beam to sweep across a target (meteorological event) M. The radar energy reflected from the target M is returned to the directional radar antenna A and forwarded to the radar receiver R. The radar receiver R processes the returned echoes RE to minimize the noise contained in this signal. The signal contains a number of noise components including, but not limited to, clutter from reflections from obstacles, topographical features in the vicinity of the radar transmitter, as well as temporal and spatial variability of the winds and precipitation, low signal-to-noise ratios, velocity folding, Radio Frequency Interference (RFI), aircraft and other point targets (e.g., birds). The radar receiver signal processor must differentiate the returned echoes RE from the noise signals contained in the received signal to accurately determine the presence, locus motion and nature of targets M that are detected in the radar antenna beam pattern.
In order to fully exploit the utility of Doppler wind measurement devices in the real time detection of meteorological phenomena, such as winds, wind shear and turbulence, it is important that the data from these Doppler wind measurement devices be quality controlled. These Doppler wind measurement devices include lidars, sodars, weather radars and clear air wind profilers. Wind profilers, for example, are able to generate Doppler moments (and hence wind and turbulence estimates) from clear-air atmospheric returns by employing a transmitter pulse generation system with a very high pulse repetition frequency (PRF) and long dwell times. In this manner, a very large number of data samples are processed per unit volume of space to increase the signal-to-noise ratio (SNR) of the return echoes. The generally weak, clear-air atmospheric signal can thus be distinguished from the ambient noise level created by the radar hardware and various environmental sources external to the radar system.
In operation, a wind profiler averages a number of data samples which comprise time domain return-power values from a given radar beam direction, then a Feast Fourier Transform (FFT) is applied to produce a single, unaveraged spectra. This step typically uses a time series of data samples collected over a half-second interval. This process for producing a single, unaveraged spectra is repeated a number of times (on the order of 60) to create a series of unaveraged spectra. The wind profiler system then subsequently averages these individual, unaveraged spectra to produce an averaged spectra The first three moments of the averaged spectra are the basic measurement data that is obtained from the wind profiler. The zero-th moment of the averaged spectra gives the total signal power and, when combined with the noise level, gives the signal-to-noise ratio (SNR) for the measurement. The first moment of the averaged spectra is the radial velocity for the given transmitted radar beam in the corresponding antenna pointing direction and, when combined with the determined radial velocities from the transmitted radar beam in other antenna pointing directions, allows for the estimation of the wind vector. The second moment of the averaged spectra provides information on the spread of radial wind velocities within the radar pulse volume and can be used to estimate wind turbulence intensities. The individual wind estimates described above are further processed via a consensus algorithm to generate a quality-controlled wind estimate. In this step, the individual wind estimates (for each altitude) collected over a predetermined time interval are examined to ascertain whether there is a “cluster” of values that all lie within a prescribed wind velocity range of each other. If such a consensus exists, the average of these values is defined as the consensus wind estimate. Otherwise, no wind estimate is reported for that altitude, for the given time interval.
Presently, wind profilers produce the consensus wind value for each altitude and this consensus value can be considered an average value, with outliner removal. A variety of contamination sources often preclude the existing consensus algorithm from producing accurate wind data. This deficiency stems from a number of sources, including: stationary and moving ground clutter, temporal and spatial variability of the winds and precipitation, low signal-to-noise ratio, Radio Frequency Interference (RFI), velocity folding, aircraft and other point targets located in the field of view of the antenna beam pattern. A number of previous attempts at solving some of these problems have been made with only limited success.
Quality control algorithms presently in use in wind profilers deal with removing contamination at the consensus wind level and hence require data gathered over an extended period of time to produce valid results. These techniques are adequate for use in human interpretation of large scale meteorological features, but are inadequate for use in automated meteorological monitoring systems to obtain timely, accurate and rapid estimates of winds, wind shear and turbul
Cornman Lawrence Bruce
Goodrich Robert Kent
Morse Corinne Sue
Duft, Graziano & Forest, P.C.
Sotomayor John B.
University Corporation For Atmospheric Research
LandOfFree
Method of moment estimation and feature extraction for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of moment estimation and feature extraction for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of moment estimation and feature extraction for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2573200