Research On Blind Separation Algorithm For Space-based AIS Received Signals

Due to the influence of ship-to-shore communication distance and communication facilities,it is difficult for the shore-based system to effectively supervise the navigation dynamics of ships in the open sea.The space-based AIS can realize long-distance monitoring of ships and master the dynamics of ships in the far sea.But the satellites have high orbits and wide coverage.There are multiple self-organized cells in the satellite observation range.The AIS signals received by the satellite receivers will inevitably collide and reduce the probability of ship detection.Therefore,this paper studies the separation algorithm of AIS conflict signals.Firstly,according to the characteristics and technical requirements of the space-based AIS system,the AIS signals received by the satellite are simulated to generate the space-based AIS signal simulation source.The cause of the space-based AIS signal collision is analyzed,and the multi-channel and the single-channel AIS collision signal are generated.Ship detection probability model,simulation analysis of conflicting signals with different degrees of separation.The simulation results show that the greater the degree of signal separation,the greater the number of ships observable by the satellite.Then,for space-based AIS conflict signals,multi-channel blind separation studies were conducted.Aiming at the problem of slow convergence of traditional EASI,a variable step EASI based on hierarchical iteration is proposed.In the initial stage of blind separation,the algorithm can achieve fast convergence,and it can achieve good steady-state performance when approaching the eost function.The simulation results show that the adaptive step size EASI algorithm and the exponential variable step size EASI algorithm,the improved EASI algorithm not only reduces the steady state error but also improves the convergence speed.Finally,considering satellite load and cost,the single-channel blind separation for space-based AIS collision signals is studied.An improved EEMD is proposed,which uses the boundary local feature scale extension method and correlation coefficient criterion to construct multi-channel,and combines the improved negative entropy-based FastICA to separate signals.The simulation results show that the proposed algorithm can achieve single-channel blind separation when multi-channel AIS signals are mixed,and the separation effect is significantly improved when compared with N-EEMD-PCA-IFastICA and EEMD-PCA-IFastICA.