Detecting Weak Wideband Signals in Interference-Dominated Environments

This paper addresses the problem of relatively wideband signal detection near the power spectral density level of the background noise from baseband data records containing multiple, stronger, narrowband interference components. Given the many unknown parameters (different relative powers, center frequencies, modulations and bandwidths, as well as time-varying aspects), many traditional interference suppression techniques such as successive interference cancellation and machine-learning algorithms are computationally intractable. As the a priori knowledge about the underlying scenario is minimal, the idea being pursued here is a more probabilistic signal-processing approach.