Impact of the Decision Function on the Overall Performance of an Energy-and Entropy-Based Hybrid Detector

The basic challenge in weak signal detection is not only to bring the signal to a useful form, but also to determine its presence in the surrounding background noise. One of the well-known and most commonly used methods is energy detection, extended in the following years to compound, hybrid solutions. In this article, the authors have shown that the improvement of detection can be achieved not only by extending already complex algorithms by additional processing branches, but simply by choosing a better decision criterion. For this purpose, the authors have defined a simulation model, which uses machine learning to redefine previously adopted decision thresholds and improve detection efficiency. The presented simulation results show significant improvement in detector performance in comparison with the classical approach.