A blind iterative calibration method for high resolution DOA estimation
Ahmed Khallaayoun, Raymond J. Weber, Yikun Huang · 2011
A novel online antenna array calibration method is presented in this paper. Unlike conventional calibration process, the proposed method is a blind approach that does not require prior knowledge of angles used for calibration and the method can be modified for real time calibration processing. An iterative process is used to converge to the actual bearings while the measured steering vector is estimated via Eigenvalue decomposition. The correction matrix is then estimated by use of the least squares estimator. The last step consists of pre-multiplying the measured data set by the obtained correction matrix which mitigates magnitude, phase, and mutual coupling effects on the acquired data set. The experimental study has been carried out on an 8-element uniform circular array (UCA) with an 8-channel RF receiver board at 5.8GHz. The performance of Bartlett, MUSIC, Beamspace MUSIC, and S2-MUSIC direction of arrival (DOA) estimation algorithms have been examined on CW and Harris SeaLancet RT1944/U radio signals. The lab test results have shown significant improvements on high resolution DOA estimation in terms of accuracy and resolution. The proposed method can be implemented in a direction finding system and can be run periodically to mitigate any changes in the environment and maintain high resolution performance over time. Potential applications of the calibration method include improvement of accurate passive RF signal source localization, high QoS wireless communications with beamforming antennas and sensor networking.