FH Sequences Selected Based on Clustering Analysis
Hua-Bin Yang, Deyu Wang · Journal of Communication and Computer · 2010
Via defining comparability of to FH (frequency hopping) sequences, mapping FH sequence into high dimension space, calculating inner product of the FH sequences, constructing appropriate kernel matrix, using cluster analysis on FH sequences set, cluster structure of characteristic space into which the FH sequences are projected as found, and then the centroid of every cluster was found out, so the FH sequence corresponding to the nearest point to centroid can be used in FH communication. The cluster analysis algorithm, such as K-means algorithm clustering and spectral clustering, minimize the distance inner cluster and maximize the distance between clusters, so the collision probability is reduced, and enhances the efficiency of FH radio network and utilize of frequency.