Detection of transient signal in additing gaussian noise: comparison between time-frequency and higher-order statistics approaches

摘要

In this paper we study two approaches for detecting transient signal in additive Gaussian noise. The traditional approach to resolve this problem is based on the energy changes into the received signal at different times. These energy detectors are usually too sensitive when the signal to noise ratio (SNR) is too weak. That is why we are interested in other algorithms using Time-Frequency transformation or Higher-Order Statistics. In the first part, these different algorithms are described and after, simulations results with artificial signals and noises are discussed for different SNR.