Abstract:

Aiming at the problem that the Gaussian component number increases rapidly with time in the density distribution fusion process, a twolevel component hybrid reduction algorithm is proposed for the different stages of the fusion process, which minimizes the loss of information. An equivalent method based on quasiMonte Carlo sampling is proposed to solve the problem that the Gaussian mixture model is no longer subject to Gaussian mixture distribution throughout the operation of Gaussian mixture exponentiation. The simulation results show that the proposed algorithm improves the fusion accuracy while ensuring the validity and feasibility of the fusion computation.