원문정보
초록
영어
As an important nonlinear filter theory, particle filter is a heated issue in domestic and foreign researches. The option of importance density is one of the key steps of particle filter algorithm. The proper option of importance density can minish the negative influence of filter algorithm caused by degeneracy problem. This paper introduces several widely-used options of importance density systemically, and analyzes their features and applied perspectives respectively. The paper also advances a comprehensive method of importance density, analyzes its technical features, explores the adjudgement and improvement of this method based on various performance, and finally puts forward the necessary further study according to the engineer requirements.
목차
1. The Optimum Importance Density Function
2. Commonly used Methods for Importance Function Generation
2.1. Use Prior Probability Density as Importance Function
2.2. Mixed Importance Function Constituted of both Prior and Posterior Probability Density
2.3. The Annealingprior Probability Density as the Importance Function
3. Use Filter Algorithm to Design the Importance Function
3.1. Use EKF to Design the Importance Function
3.2. Design the Importance Function through UKF
3.3. Design the Importance Function through Gaussian-Hermite Filter (GHF)
4. Mixing of Prior Probability Density and UKF to Generate Importance Function
5. Conclusion
Acknowledgements
Reference