Adaptive Calibration for Multi-Channel SAR Instruments

Kurzfassung

In conventional Synthetic Aperture Radar (SAR) instruments the azimuth resolution as well as the coverage (swath width) is inversely proportional to the Pulse Repetition Frequency (PRF), in order to avoid ambiguities. By employing multiple channels in azimuth and elevation in combination with Digital Beam-Forming (DBF) techniques the contradicting requirements to achieve a wide coverage and fine resolution can be overcome. A basic DBF technique in elevation is Scan-On-Receive (SCORE). A time-varying high-gain receive beam is directed to the position of the SAR signal echo on the ground. The narrow receive beam additionally leads to a better suppression of range ambiguities and a higher Signal-to-Noise Ratio (SNR). The DBF pattern is formed by combining the weighted channels. Errors in the individual channels, such as imbalances, drifts, coupling, noise, gain compression, changes in the frequency transfer function or ADC errors (aliasing, quantization and jitter), where the influence of the error sources can vary from channel to channel, lead to different influences on the amplitudes and phases of the received signals. Thus, the combination might result in a degradation of the beamformed pattern and therefore of the performance, e.g. worsened SNR, decrease resolution, higher ambiguity levels and degraded polarization. Giving rise to the need of a proper calibration of the channels. Due to the real-time constraints imposed by SCORE the calibration has to be performed in real-time on board of the satellite. An internal calibration for active phased arrays can be executed by employing a single tone calibration signal that is coupled in close to the antenna and evaluated at the digital beamforming units. It allows for a calibration simultaneous to the radar operation. Additionally, the individual elements of the channels can be calibrated by altering where the calibration signal is coupled in. In order to account for frequency variations, the frequency of the calibration signal has to be varied. The calibration sequence determines the order, on pulse-to-pulse basis, in which the elements are characterized, at which frequency and for which calibration loop (e.g. receive, transmit and short). Conventionally, the characterizations alter regularly and the sequence is fixed. However, the accuracy of the characterizations might vary. For example, it depends on the frequency of the calibration signal with respect to the frequency band of the SAR echo signal (inband vs. out-of-band). Further, the drift in some elements might be stronger than in others (inner vs. outer elements). The result is an increase of the overall residual error if all parameters in the calibration sequence are treated equally. Here, a mathematical model to describe the effects in a multi-channel SAR instrument as well as a simulation tool based on said model is introduced. To mitigate the residual errors, an adaptive calibration method is proposed. It alters the calibration sequence on-the-fly based on the instrument’s behaviour, e.g. increase in-band characterizations. Thus, reducing the residual error. Simulation results showing the performance of the proposed method are presented.