Spectrum sensing based on goodness of fit tests

摘要

This paper applies three methods of goodness of fit test, namely Anderson-Darling (AD) test, Cramer-Von Mises (CM) test and Kolmogorov-Smirnov (KS) test, into spectrum sensing in a cognitive radio. A lower bound on detection probability on each application is derived, and simulations in different environments are given. These results show that performances of three tests above are all better than the traditional energy detection under environment of AWGN channels and slow, flat fading channels.