Study on intelligent data assimilation method based on Equivalent-Weights Particle Filter

In the study of ocean dynamics, the particle filter method has gradually become a research hotspot in the field of data assimilation due to its advantages in nonlinear and non-Gaussian systems. At present, the most direct sources of information on marine environmental elements are buoys and autonomous unmanned observation platforms, but the observation data obtained by these two methods are sparse and unevenly distributed. Therefore, how to use the particle filter method to organically integrate the sparse and unevenly distributed observation data with the ocean numerical model to improve the accuracy of analysis and forecasting of marine environmental elements is a difficult problem that needs to be solved urgently. Based on the equivalent-weights particle filter method and the characteristics of the particle swarm optimization algorithm that can search for the global optimal solution, this paper studies and analyzes the defects of the traditional equivalent-weights particle filter method that relies on future observation information to adjust the particles, and proposes an equivalent-weights particle filter assimilation method based on the particle swarm optimization algorithm. The proposed method is experimentally verified for sparse and unevenly distributed marine environmental observation data combined with regional ocean numerical models.