Abstract:

Conventional MUSIC algorithm is hard to be applied in MBBS for its great computational load, and most of the beamspace MUSIC algorithms need the estimation of the covariance matrix and eigenvalue decomposition, which leads to complex signal processing platform. To solve this problem, a MM-MUSIC algorithm is proposed on the combination of the fast subspace estimation based on MSWF and a MSB-MUSIC algorithm. Compared with MSB-MUSIC algorithm, the proposed MM-MUSIC algorithm has much fewer computational load and higher parallelism at a cost of a little loss of performance. The realization of fast subspace estimation with AccelDSP and the processing results of experiment data prove the efficiency and practicality of the MM-MUSIC algorithm.