ON ADAPTIVE MULTIBAND SIGNAL-DETECTION WITH THE SMI ALGORITHM

In a nonstationary and/or nonhomogeneous interference environment, an adaptive system for target detection may suffer a severe performance degradation, due to the lack of a sufficient amount of data from which the system can “learn” (estimate) the statistics of the environment. We analyze the detection performance of an adaptive system, which employs a frequency diversity (multiband) signaling waveform and a multiband sample matrix inversion algorithm (SMI). By comparison with the corresponding single-band system under the chosen system constraint, we show that the multiband system can significantly outperform the single band when the amount of data available from a single frequency band is severely limited by the environment. © 1990 IEEE