Improving the Performance of the Capon Algorithm by Nulling Elements of an Inverse Covariance Matrix

It is well known that the Capon algorithm offers better direction of arrival (DoA) estimation performance compared to the Fourier method (FM). This is due to the fact that, unlike in the FM, the total output power of antenna array is minimized while maintaining a constant gain at a certain look direction in the Capon algorithm. Unfortunately, the DoA estimation performance of the Capon algorithm is degraded as the signal-to-noise ratio of received signal is lowered and thus, it often fails to distinguish incident signals having similar DoAs. In this paper, we propose a scheme improving the DoA estimation performance of the Capon algorithm by nulling certain elements of an inverse covariance matrix used in the Capon algorithm.