An ICA Algorithm Applied To Signal Deinterleaving

Signal deinterleaving means deinterleaving the same emitter signals in the mixed signal observed into the same category. With the increaseing complexity of the electromagnetic environment, the traditional signal deinterleaving algorithms depending on generic parameters have turned out to be unseasonable due to the pulse aliasing, the high correlation among the parameters and the complex modulation modes.Taking the above issues into consideration, this paper has described a signal deinterleaving algorithm based on Independent Component Analysis(ICA). The biggest advantage of the algorithm is its hardly needing any prior knowledge of the source signals. The main work of the paper shows below.The paper summarizes the traditional signal deinterleaving algorithms and point out their limitations and where the limitations come from.And then the paper summarizes the Blind Source Separation(BSS) algorithm and its appropriateness for signal deinterleaving. Also a FastICA algorithm has been introduced into signal deinterleaving. First a signal-processing using the normalization method and whitening processing is apllied with the intercrepted signal before the independent component analysis process. Then the target signal can be got by combining the signals with same parameters from ICA processing over different times. The method is verified correct and better than the traditional by simulation.The single channel receiver is more commonly found in the practical uses and a signal deinterleaving algorithm combining Empirical Mode Decomposition(EMD) and ICA is put forward. Simply the method is to decompose the signal observed into several Intrinsic Mode Function(IMF) and get them through ICA separation. The method is verified correct and useful for the frequency agile radar signal deinterleaving by the simulation.