Techniques for Calibrating Numerically Modeled Array Manifolds, 16-9220 Printer Friendly VersionPrincipal Investigator Inclusive Dates: 10/01/01 - Current Background - Radio direction finding (DF) is a process that requires prior knowledge of the installed DF antenna patterns as a basis for estimating direction of arrival for a signal of interest. Although SwRI’s numerical electromagnetic models are increasingly accurate in their predictions of the individual installed antenna patterns, residual errors do remain in the predicted patterns. These models do not address the influences of circuitry internal to the DF system. These internal influences can introduce azimuth-independent offsets that, if left uncompensated, render the electromagnetic models unusable for DF. A data-efficient adjustment procedure is needed that will account for azimuth independent differences and both azimuth- and frequency-dependent differences between the numerically modeled and actual responses. Approach - The approach for achieving this objective is based on a model that assumes that the numerically generated response vectors differ from the actual system response vectors because of two basic factors. These two factors are the azimuth-independent antenna offsets and the azimuth-dependent pattern differences. Under this model, estimates of the system response vectors will be recovered from the numerical response vectors by successive removal of the azimuth-independent offsets and the azimuth-dependent differences. To achieve significant savings in the required volume of calibration measurements, frequency regression and azimuth/frequency regression techniques will be developed to exploit both the azimuth and frequency dependence of the observed differences. Accomplishments - A new calibration-processing scheme has been developed with the important advantage of requiring a minimum number of calibration measurements, while taking full advantage of all additional calibration measurements beyond the minimum requirement. The immediate advantage of the new process is the opportunity to correspondingly reduce the calibration time by cycling through a much longer list of calibration frequencies as the ship is turning at a fixed rate through each calibration circle. |