Modified Gini Index Detector for Cooperative Spectrum Sensing over Line-of-Sight Channels

Recently, the Gini index detector (GID) has been proposed as an alternative for data-fusion cooperative spectrum sensing, being mostly suitable for channels with line-of-sight or dominant multi-path components. The GID is quite robust against time-varying noise and signal powers, has the constant false-alarm rate property, can outperform many the state-of-the-art robust detectors, and is one of the simplest detectors developed so far. The modified GID (mGID) is devised in this article. It inherits the attractive attributes of the GID, yet with a computational cost far below the GID. Specifically, the time complexity of the mGID obeys approximately the same run-time growth rate of the GID, but has a constant factor approximately 23.4 times smaller. Equivalently, the mGID takes approximately 4% of the computation time spent to calculate the GID test statistic, which brings a huge reduction in the latency of the spectrum sensing process. Moreover, this latency reduction comes with no performance loss with respect to the GID.

Automated summary

Recently, the modified Gini index detector (mGID) has been proposed as a more efficient alternative to the Gini index detector (GID) for data-fusion cooperative spectrum sensing in certain types of channels. The mGID inherits the robustness and simplicity of the GID while significantly reducing the computational cost and latency. The mGID achieves approximately 23.4 times smaller constant factor in time complexity, resulting in a 4% computation time compared to the GID test statistic calculation with no loss in performance.