Joint State and Fault Estimation for Nonlinear Systems Subject to Measurement Censoring and Missing Measurements

Area of Science:

• Control Systems Engineering
• Nonlinear System Analysis
• Signal Processing

Background:

• State estimation for nonlinear systems is challenging due to measurement censoring (MC) and missing measurements (MMs).
• Actuator faults further complicate accurate state estimation in dynamic systems.
• Existing methods often struggle with combined MC and MMs, necessitating advanced estimation techniques.

Purpose of the Study:

• To develop a novel joint state and fault estimation framework for nonlinear systems.
• To address the challenges posed by simultaneous measurement censoring and random missing measurements.
• To enable robust and simultaneous estimation of system states and actuator fault signals.

Main Methods:

• Integration of an improved Tobit Kalman filter to model measurement censoring.
• Utilization of Bernoulli random variables to characterize random missing measurements.
• Application of a federated fusion approach for distributed estimation and centralized fusion.
• Consideration of abrupt and ramp actuator faults within the estimation framework.

Main Results:

• A novel joint estimation framework is proposed, combining Tobit Kalman filtering and federated fusion.
• The framework achieves simultaneous robust estimation of system states and fault signals.
• The boundedness of the filtering error for the designed estimator is proven under specific conditions.
• Effectiveness demonstrated through two engineering experiments.

Conclusions:

• The proposed framework effectively handles nonlinear systems with both measurement censoring and missing measurements.
• The integration of Tobit filtering and federated fusion provides a robust solution for joint state and fault estimation.
• The method offers improved reliability for state estimation in complex and noisy environments.