Performance bounds on cyclostationary feature detection over fading channels

Cyclostationary feature detection enables a cognitive radio system to reliably sensing the existence of licensed users. A statistical test based on the second-order cyclostationary features has been widely used. This paper evaluates the detection performance of this 2^nd-order feature based technique subject to Nakagami fading. The analytical upper and lower bounds on the detection probability are derived by exploiting some available series expansion and exponential-type bound of the generalized Marcum Q-function. As a result, the preliminary detection performance can be acquired without relying on Monte Carlo simulations.