Variable forgetting factor PASTd algorithm for time-varying subspace estimation

A new subspace tracking algorithm based on the PASTd algorithm is proposed. The algorithm is obtained by introducing the variable forgetting factor, w hich adapts itself to the time-varying subspace environments. The tracking capability of the proposed algorithm is demonstrated by computer simulation s in an abruptly changing DOA scenario. The estimation results of the varia ble forgetting factor PASTd (VFF-PASTd) algorithm outperform those of the P ASTd algorithm in the nonstationary case as well as in the stationary case.