Simple and practical cyclostationary beamforming algorithms

Inspired by the asymmetric principal component analysis (APCA) neural model and based on the signal cyclostationarity, the authors propose a simple gradient-descent beamforming (GDB) algorithm. The GDB algorithm is an adaptive blind beamforming algorithm and can be used to extract signals with cyclostationarity under a complex signal environment. Although the GDB algorithm suffers from slow convergence, it has a low computational complexity. Two more algorithms, called the boosted GDB and the beta GDB algorithms have been defined based on the GDB algorithm. All the algorithms have been simulated and compared. The boosted GDB and beta GDB algorithms are shown capable of providing fast convergence and satisfactory signal-to-interference-and-noise ratio (SINR) performance, and can be used for implementation in real-time systems.