Adaptive DOA tracking by rank-revealing QR updating and exponential sliding windows techniques

It is well known that adaptive eigensubspace methods have been developed for tracking time-varying directions-of-arrival. Nevertheless, they involve a high computational burden and are not easily mapped onto VLSI implementation. To alleviate this problem, a more efficient noise-subspace updating based directly on an exponentially sliding data matrix and its rank-revealing QR decomposition (RR-QRD) is suggested. Two algorithms for updating the noise-subspace is proposed by (1) updating the Q factor, and (2) updating R factor of the RR-QRD, respectively. Besides the computational and implementational benefits, simulation results show that both the proposed methods have good tracking performance