Error Analysis of an Airborne FMCW ATI SAR for Ocean Surface Currents Retrieval

Abstract

Along-track interferometric synthetic aperture radar (ATI-SAR) has shown its promise in high-resolution ocean surface currents velocity mapping. An accurate velocity measurement by ATI-SAR requires precise phase difference or interferometric phase estimation between receivers. This work explores possible instrumental and platform error sources that may introduce the observed ATI phase error from field experiments. With technical analysis and literature review, a number of possible error sources were presented and a comprehensive error analysis was performed to study the effects of the phase mismatch between receiver channels, phase imbalance between receive antennas, uncompensated motion errors due to the inaccurate measurements of platform attitude/velocity, and ATI phase errors due to the uncertainty of the reference digital elevation model (DEM) map. The range-dependent phase fluctuations observed in ATI-SAR interferogram are demonstrated to be caused by the mismatch in the phase responses between the two receiver channels of the radar. A novel phase ripple calibration scheme was presented to mitigate the phase fluctuations during the SAR processing and could be applied to real-time calibration. The residual phase offset in the range direction is proved to be most likely due to the imbalance between the receive antennas’ phase patterns. The estimation of the antenna phase imbalance with the collected farmland data was applied to remove most of the residual phase offset. The phase undulations observed in the along-track direction of the interferogram are demonstrated to be related to the uncompensated motion errors introduced by the inaccurate attitude and velocity measurements by the inertial navigation system (INS). The DEM uncertainty analysis shows that the ATI phase error due to the DEM height error cannot be neglected when the DEM uncertainty exceeds a threshold. With all proposed calibration approaches, the ATI phase error can be reduced by as much as 1 radians after calibration, which converts to about 55 cm/s ocean surface velocity error. Assuming no ocean waves exists in the imaged scene, which means the Doppler measurements made by ATI-SAR consist of contributions only from the ocean surface currents, the estimated surface velocity error can be reduced from 70 cm/s to about 14 cm/s, which is a significant improvement for ocean surface currents mapping with ATI-SAR.