Linear prediction modeling for signal selective DOA estimation based on higher-order statistics

The direction-finding approach for impinging signals is one of the most important issues in array processing. By exploiting the cyclic statistics and higher-order temporal properties of communication signals, cyclic higher-order statistics (CHOS) direction-finding approaches have been proposed for narrow-band non-Gaussian signals. However, conventional cumulant-based algorithms become very complicated and are computationally intensive when a cumulant higher than the forth-order is used. In this paper, by utilizing a linear prediction (LP) model of the sensor outputs, a new cyclic higher-order method is given to detect the signals of interest (SOI). The proposed method can not only reduce the computational load and completely exploit the CHOS temporal information, but can also correctly estimate the DOA of desired signals by suppressing undesired signals. We also show the effectiveness of the proposed method through simulation results.