Precise Incident Power Density Pattern Measurement through Antenna Pattern Deconvolution

Future cellular and fixed wireless loops (FWL) systems employing highlydirective base stations antennas (5 to 8 degrees beamwidth) and moderatelydirective subscribers' antennas (15 to 25 degrees beamwidth) offer very largecapacity due to reduced interference. An important property of the environmentof such systems is the interference caused by scattering of the signal froma subscriber transmitter into directions other than the direct line of sightbetween the subscriber and the base station. In particular, for multibeam basestation applications scattering could arrive at the base station in thedirection being used by a different beam, resulting in interference that isdifficult to reduce by normal nulling techniques. Thus interference can behighly dependent on the incident power density pattern (IPDP) caused byscattering of the signal radiated from the subscriber. We discuss herein theuse of a uniformly illuminated array accompanied by electric fielddeconvolution to measure the crucial IPDP with the same performance as a lowsidelobe array of the same size. The mathematical correction technique usesdeconvolution of the measured complex electric field pattern with that of theantenna in free space by means of the Fourier Series and limiting the rangeof Fourier coefficients to those that are not negligible in the free spacepattern. Application of the technique to an experimental uniform array witha 2.5 degree azimuthal beamwidth shows the practicality of the deconvolutionwith real antennas in real environments. The improved resolution and accuracyprovided by Taylor weighting versus unweighted deconvolution when trying tomeasure weak scattered components in the presence of a nearby strong specularcomponent is demonstrated. The IPDP was measured from many sites surroundinga suburban base station. A plot of the cumulative distribution of the ratioof the widely scattered power to that within a prescribed beamwidth summarizesthe result of using the deconvolution technique on this experimental data.