Realization and comparison of QRD algorithms for STAP

For space time adaptive processing of airborne multichannel array radar, the computation load of calculating weighting vector is the heaviest. With sample matrix inversion (SMI) algorithm, QR decomposition usually is used to calculate weighting vectors. Four types of QR decomposition algorithm, Gram-Schmidt, modified Gram-Schmidt, Householder transform and Givens rotation, were studied in this paper. Through realization with ADSP-TS201, the computation efficiency and numerical stability of these four algorithms were compared. The realization results were validated with simulated multi-channel echo data.