A Recursive Estimation Algorithm for Discrete-Time Systems with Unknown Noise Parameters.
Abstract
The purpose of this report is to develop a recursive algorithm for state estimation for systems with uncertain models. The algorithm uses a combined detection-estimation scheme whereby the set in which the uncertainties are contained is detected, and appropriate estimator is then used. The approach used is an extension of a weighted minimax performance criteria to the dynamic case. Since global optimal solution is not possible, an approximate algorithm is derived which only optimizes the stage-by-stage performance without changing any past decisions. The expressions for the algorithm and an approximation of its performance are derived. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1978
- Accession Number
- ADA069794
Entities
People
- Nathan Guedalia
Organizations
- University of Illinois Urbana–Champaign