Independent Component Analysis Of Gravity Tide Signal Based On Improved PSO And Tidal Harmonic Extraction

Gravity earth tide contains abundant harmonic informations which may infer the precursor information of the earthquake, so how to get the harmonic informations of the gravity earth tide is a key problem. Based on the generating mechanism, this paper introduces a model to decompose the signal which solves the problem that the traditional method can’t classified the harmonic into series.Because of the model is a theoretical model, so this paper is to seek for an algorithm separating the signal into parts which is according to decomposition model.Independent component analysis (abbr.ICA) is a method to seperate the blind source separation (abbr.BSS) problem. Because the gravity earth tide is a mixed signal, so ICA is applied to separate gravity tide signal. ICA has two parts, objective function and optimization algorithm. Based on the difference of the optimization algorithm, ICA is divided into different kinds. Traditional ICA exist the problem which is depended on the gradient information. So this paper uses the particle swarm optimization algorithm to obtain better separation matrix.Particle swarm optimization algorithm (abbr.PSO) is built according to the behavior of the flying birds to seek for the food. The algorithm has low complexity, so the paper uses the PSO to optimize the objective function of ICA. In view of the particle is easy falling into local optimum particle, this paper puts forward an improved PSO which improves the convergence speed of the algorithm and could achieve the global optimal particle.In the end, the paper uses the method which is based on the improved PSO and ICA to separate gravity earth tide. The results show that the paper could get the component which is consistent with the decomposition model. At the same time, the spectrum analysis of the signal shows that the signal is separated into three parts which is according to the model. So the model which is improved in the paper is an effective decomposition model and the proposed algorithm in the paper is also a kind of effective method to decompose the signal of the gravity earth tide.