A novel space-time multi-user detection algorithm was proposed

which combined a power of covariance matrix R technique with MOE blind adaptive space-time multi-user detection. Moreover

the performance of the improved MOE blind space-time multi-user detector was analyzed

according to the effect of white noise

data sample number and positive integer u in cost function. The theoretic analysis and simulation results show that the performance of the novel multi-user detection method asymptotically converge to the MMSE blind space-time multi-user detector on condition that smaller noise variance

larger enough data sample number and u → ∞ .