Innovation generation for bilinear systems with unknown inputs
The paper addresses the problem of filter design for fault detection in discrete-time bilinear systems. First a uniformly completely observable subsystem without unknown inputs is extracted from the original system. Next, a Kalman filter is designed for the bilinear system resulting from the first step. The innovation in the presence of additive faults is discussed, and the generalized likelihood ratio (GLR) test provides an estimate of the fault magnitude and the fault occurrence time. Finally, the theory is applied to design and validate a fault detection and isolation system for a three-tank process. For the entire collection see [Zbl 0915.00063].
- Innovations generation in the presence of unknown inputs: Application to robust failure detection
- Multi-innovation identification methods for linear-parameter systems
- Innovations system representation in the optimal LQG regulation problem
- Multi-innovation identification methods for input nonlinear equation-error systems
- An innovation representation for nonlinear systems with application to parameter and state estimation
- Multi-innovation identification methods for output-error systems
- Decomposition-based multiinnovation gradient identification algorithms for a special bilinear system based on its input-output representation
- New method for predictive controller design for bilinear systems
- scientific article; zbMATH DE number 5499506
- Innovation systems by nonlinear networks
This page was built for publication: Innovation generation for bilinear systems with unknown inputs
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4256898)
