Evaluating Two Array Autocalibration Methods with Multifrequency HF Radar Current Measurements

Abstract

ne pivotal factor affecting the accuracy of HF radar current measurements is the direction of arrival (DOA) estimation performance of the current signal. The beamforming technology or superresolution algorithm cannot always perform best in practical applications because of the phase errors existing in array channels. These phase errors, which cause uncertain estimation of DOA, lead to confused values in radial current maps. To solve this problem, this paper is focused on discussing the performances of two autocalibration methods using sea echoes for multifrequency high-frequency (MHF) radar current measurements. These two array calibration methods, based on maximum likelihood (ML) and multiple signal classification (MU), first seek single-DOA sea echoes and then gather them for array calibration using different cost functions. The ML and MU methods provide approximate mean phases, while the standard phase errors of the MU method are smaller. After array calibration using these two methods, the results show significant improvements in current retrievals. Comparisons between the MHF radar and ADCPs reveal that array calibration using the ML and MU methods also improves the estimation of radial currents clearly, with correlation coefficients over 0.93 and rms differences of 0.09?0.18 m s?1 at different operating frequencies and sampling locations. The performance of the bearing offset is also improved. Only small bearing offsets less than 10° exist in radial current measurements. Therefore, this paper demonstrates that array calibration is a crucial part for current measurements, especially for direction-finding HF radar.