Parameter Optimization for the Detector Based on Fractional Lower Order Moments in Alpha-Stable Distributed Noise

In impulsive noise modeled by the symmetric alpha stable (\(S\alpha S\)) distribution, the detector based on fractional lower order moments (FLOM) is a good choice for signal detection due to its simplicity and satisfied performance. The moment order is a priori parameter of the FLOM detector. However, there is no quantitative result to help us to choose a good moment order in the literature. In order to calculate the optimal value of the moment order for the FLOM detector, we first derive a novel asymptotic expression for the detection probability and then calculate the optimal value of the moment order by the Genetic algorithm. The effect of different parameters on the optimal moment order is discussed and some useful conclusions are drawn. Simulation results prove that the optimized moment order makes the FLOM detector achieve the best performance in different \(S\alpha S\) noise.