The performance of spectrum sensing algorithm based on goodness of fitting（ GOF） testing is excellent

however

its complexity is high. New detecting model is given

introduce generalized likelihood ratio

the square of sample mean of received signal is employed to construct new test statistics; probability density function（ PDF） of test statistic and the expression of decision threshold are deduced

so as to achieve spectrum sensing. Compare GOF algorithm with energy detection（ ED） algorithm in Gaussian channel

simulation show that the proposed algorithm has good detecting performances and has low complexity.