The research and FPGA implementation of signal subspace decomposition

The decomposition of signal subspace and noise subspace is a difficult problem for hardware implementation of MUSIC algorithm. In order to solve the above problem, this paper researches multiple Jacobi algorithms and adopts the combination of the sorted and clearance Jacobi algorithm to improve the efficiency. Meanwhile, this paper gives some algebra for source number estimation. Compared with traditional information theory method, the new method can reduce the computing complexity of estimation of source number efficiently. Finally, this paper achieves a new subspace decomposition architecture based on FPGA, which solves eigenvalue of an 8*8 matrix in 86.83 us and estimation of source number in 24.66 us. The FPGA verification results show that the presented method can decompose signal subspace with high accuracy (up to 10^-4) and a good compromise in area and speed.