System modeling based on orthotopic expansion particle filter for linear time-varying system with bounded noise

To make parameter estimation more accurate and concise, a parameter identification method based on particle evaluation and orthotope is presented in this paper. First, all possible expansion directions of the parameter-feasible domain (predicted orthotope) were evaluated. Further particle evaluation was carried out in the expanded orthotope to shrink the feasible domain. Once the upper and lower bounds of the parameter estimation were obtained, a particle filter with improved sampling was used for state estimation. Finally, the effectiveness and accuracy of the proposed algorithm were verified through simulations.