Credibility test for blind processing results of sinusoid using Chebyshev's Inequality

Abstract

Blind processing of the intercepted signal, including modulation recognition and parameter estimation, is a classical issue for cognitive radio (CR) and electronic intelligence (Elint) applications. In the non-cooperative condition, it is essential that each processing result be accompanied by a reliability estimate or credibility metric. There is extensive literature regarding the subject of modulation recognition, parameter estimation as well as their accuracy evaluation by analysis or simulation. But, as far as we know, none of them have dealt with the credibility test on the recognition and parameter estimation results for an each individual processing. In this paper, the test is modeled as a hypothesis which aims at deciding whether an individual modulation is correct or not and whether specific parameter estimation is accurate or inaccurate. The null hypothesis is that an individual modulation recognition is correct and the parameter estimation is accurate. The credibility test is performed by using the Chebyshev's Inequality (CI). Simulation results show that the proposed method is superior for credibility test of blind signal processing for sinusoid even at low signal-to-noise ratio.