To overcome the limitations of energy detection which cannot differentiate primary user（PU）from secondary user（SU）

blind source separation algorithm based on fast independent component analysis（ICA）is introduced to separate the mixed signals of PU and SU.A decision rule based on approximation of negentropy is proposed to perform detection after separation.Then a weighted cooperative sensing method based on fast ICA is proposed to combat the unreliability of single node sensing.Finally

a selective cooperative sensing algorithm is also proposed to reduce the transmission overhead and detecting delay.Simulations results show that the proposed algorithm can realize reliable detection of PU and increase the spectrum efficiency.