원문정보
초록
영어
This paper discusses a new method of HIMM for the target tracking problem. This new method is a combination of the H∞ filter and the interactive multiple model (IMM) algorithm. The sub-filters of IMM algorithm are based on Kalman filter, in which the noise statistical characteristic should been known exactly, and in fact we only can get its approximatively models that worsen the performance of IMM algorithm. However, H∞ filter is not need to carry on any supposition to the noise statistical property, which is not sensitive to the noise, and has robustness to the uncertainty of noise. A meaningful example is presented to illustrate the effectiveness of the authors’ method, the performances of IMM and HIMM in terms of stability, tracking accuracy and robustness are compared. The purpose of this paper is to demonstrate the effectiveness of applying the HIMM algorithm on the target tracking problem, which in the past have typically been solved by Kalman filters.
목차
1. Introduction
2. H Robust Filtering Algorithm
2.1. Discrete H filtering
2.2. Suboptimal H filtering
2.3. Comparison of H Filter and Kalman Filter
3. IMM Algorithm
3.1. The Principle Of IMM Algorithm
3.2. The Steps of IMM Algorithm
4. HIMM Algorithm
5. Simulation Results
6. Conclusions
Acknowledgements
References