Communications: directive radio wave systems and devices (e.g. – Testing or calibrating of radar system – By monitoring
Reexamination Certificate
2002-08-14
2004-04-13
Gregory, Bernarr E. (Department: 3662)
Communications: directive radio wave systems and devices (e.g.,
Testing or calibrating of radar system
By monitoring
C342S156000, C342S173000, C342S195000
Reexamination Certificate
active
06720911
ABSTRACT:
FIELD OF INVENTION
This invention relates to calibration of arrays of direction finding antennas and more particularly to antenna array calibration using a reduced set of sampling frequencies.
BACKGROUND OF THE INVENTION
The inherent accuracy of correlation interferometer direction finding, CIDF, algorithms is described by characterizing all the direction finding antenna array responses in terms of amplitude, and phase or I&Q. In order to characterize the array, one develops a system resident, direction finding, DF database array manifold that contains the antenna responses. This characterization process is called calibration. Calibration data is measured and recorded over many frequencies and azimuths for selected elevation/depression angles and polarizations depending on the DF system requirements in order to produce a DF manifold or database.
Array manifold development is generally obtained by measuring and recording antenna data from the DF array installed on the ship, editing the data to filter out bad measurements, interpolating the data to produce the required array manifold azimuth distribution, and formatting the data for system installation.
In the past, there were two test measurements that were used to generate the array manifold. The first measurement was a surface wave measurement which involved measuring the antenna array response to signals originating from a known location that follow the curvature of the earth from on-shore. The second set of calibration values came from sky wave measurements which measure the response of the array to signals that are reflected by the ionosphere down to the antenna array. Since it is impractical to measure sky wave response on a ship, calibration, in the past, has involved the use of brass ship models which are {fraction (1/48)}th scale models of the ships in question. In order to do the calibration, the scale model is provided with loop type antennas at the positions about the ship that mimic those for the full-scale platform.
The scale model is then rotated relative to a calibration source so as to provide the surface wave and sky wave data. These measurements are taken at all frequencies of interest, which typically number 90 in a band from 1 MHz to 30 MHz.
There is, however, a difference between the results from the model and those from the full-scale platform or ship. In order to adjust the results from the model to match those of the ship, a complex optimization procedure is used which involves generating complex weights that are used to correct the response of the scale model array so that it corresponds to the shipboard array.
Key to making sure that direction-finding using such an array is effective, calibration procedures involve the selection of a number of frequencies at which signals are expected to arrive. As mentioned above, typically 90 such frequencies cover a range of 1 MHz and 30 MHz. It will be appreciated that with respect to surface wave calibration, the calibration antenna must transmit signals at these 90 frequencies to the ship, and the response of the shipboard antenna array is measured for each of these frequencies at 2° increments in azimuth angle so that a database called an array manifold is created.
In order for all 90 frequencies to be sampled at 360 different azimuth angles, a ship can be on station for as long as 36 hours. To obtain the required data, the ship must steam in circles, for instance, ten miles away from a shore station that is transmitting calibration signals so that the response of the shipborne array can be measured. These measurements are then used with surface wave and sky wave measurements from the scale model to generate an array manifold or database. Signals from the antenna array aboard the ship are subsequently used to generate complex weights which will adjust the model array response such that when these normalized signals are passed to a direction finding algorithm, the result is an accurate line of bearing angle between the ship and the electromagnetic source.
Note that once the surface wave components of the calibration have been generated at sea, the resulting characterization array is used in combination with the scale model to generate the weights for a full complement of frequencies for both surface waves and sky waves. Thus, having characterized the shipboard array and having data from the corresponding scale model, it is then possible to calculate weights which are to be applied to the model array data so as to correct them or normalize them to be able to be used by the direction finding system on the ship.
The problem with this calibration method is that getting the surface wave measurements at sea takes a long time. This is because ship must be rotated through a full 360° for each sampling frequency.
SUMMARY OF THE INVENTION
In order to eliminate having to sample all frequencies which are deemed to be important on a shipboard calibration run, it has been found that the weights associated with selected frequencies are valid across a wide bandwidth. This means that weights generated at one frequency are valid for a number of adjacent frequencies. As a result, the number of frequencies involved in the shipboard calibration phase can be reduced by as much as 80%.
Were the bandwidth of the complex weights at all frequencies uniform and linear, then all that would be necessary would be to simply divide up the total bandwidth into a convenient subset of equally-spaced frequencies. However, due to scattering and edge effects as well as other topographical artifacts associated with the ship, during the initial placement of the antennas, a frequency regime is set up in which the frequencies at which calibration measurements are specified. These frequencies are more dense in some regions, whereas in other regions, there is a less dense number of calibration frequencies.
It has been found that the complex weights to be specified in the manifold or dataset are useful over a broad range of frequencies. Thus, weights for one frequency will work over a number of adjacent frequencies. As a result, one can get broadband performance out of a set of complex weights. Whereby the complex weights need be provided only for a reduced set of the originally specified calibration frequencies.
For shipboard calibration measurement in order to reduce the total number of frequencies to about 15-20% of the original number, one must be careful not to remove too many frequencies in regions where the density of calibration frequencies is high. The density of calibration frequencies is high when there are resonances causing the antenna patterns to change very quickly. Here it is necessary to provide a larger number of calibration frequencies, although still less than the original number of calibration frequencies determined from a scale model of the ship. Where there are less dense frequencies even less numbers of calibration frequencies need to be provided. What this means is that one is culling the calibration frequencies at which one wants to sample and eliminates those frequencies based on the characteristics of the antenna array and the topside configuration of the ship.
As to what calibration frequencies are initially picked, and from which one can ascertain frequency density, the frequencies picked relate to direction finding error based on the results of direction finding using the original model. This results in a set of calibration frequencies.
How this is done is as follows: Considering a frequency of interest that is halfway between two frequencies, there is going to be a direction finding error associated with it because one is not on the calibration frequency anymore. How far a target frequency can be from a calibration frequency and still be within an acceptable DF error defines band edge error. When the band edge error is above some limit, one needs to add another calibration frequency. Thus band edge error establishes frequency density.
It has been found that more dense calibration frequency densities occur when antenna array patterns are changing quite rapidly with frequency
BAE Systems Information and Electronic Systems Integration Inc.
Long Daniel J.
LandOfFree
Method and apparatus for reducing the amount of... 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 and apparatus for reducing the amount of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for reducing the amount of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3223161