Nonlinear Estimation Framework for Optimized Target Tracking, Estimation, and Self assessment
Abstract
The research focused on advancing the theory of multi-dimensional non-linear, non-Gaussian state estimation methods with special attention on monitoring and ensuring credibility (also called consistency) of the estimates produced by the estimation algorithm. Estimate consistency is important in safety-critical applications and is often required by the user. For the purpose of monitoring consistency of estimates produced by the current estimation algorithms, two methods were designed: the cooperative filter design combining estimates from two filters to provide the consistency information and the entropy-based consistency monitoring designed especially forthe stochastic integration filter to provide the consistency information.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 05, 2020
- Accession Number
- AD1106331
Entities
People
- Ondřej Straka
Organizations
- University of West Bohemia