A Variable Structure Multiple-Model Estimation Algorithm Aided by Center Scaling

The accuracy for target tracking using a conventional interacting multiple-model algorithm (IMM) is limited. In this paper, new variable structure of (VSIMM) aided by center scaling (VSIMM-CS) proposed to solve problem. novel VSIMM-CS has two main steps. Firstly, we estimate the approximate location true model. This expected-mode augmentation (EMA), and method—namely, expected model optimization method—is further enhance EMA. Secondly, change original set ensure current as symmetry set, scaled down certain percentage. Considering linearity system, errors produced symmetrical models can be well offset. Furthermore, narrowing distance between default another effective method reduce error. second step based on theories: symmetric proportional reduction method. All theories aim minimize much possible, simulation results highlight correctness effectiveness methods.