Fault-Tolerant Cubature Kalman Filter for Engineering Estimation Control Systems

Area of Science:

• Control Systems Engineering
• Signal Processing
• Nonlinear System Analysis

Background:

• The cubature Kalman filter (CKF) is effective for nonlinear systems but struggles with abnormal sensor measurements.
• Abnormal measurements can alter measurement noise covariance, degrading filter performance and potentially causing failure.
• Fault-tolerant filters address abnormal measurements, but theoretical analysis is limited.

Purpose of the Study:

• To establish a measurement noise model and analyze noise deviation magnitude.
• To develop and analyze a fault-tolerant cubature Kalman filter (CKF) for systems with abnormal measurements.
• To propose an improved fault detection algorithm addressing CKF sampling approximation deviations.

Main Methods:

• Established a measurement noise model and analyzed noise deviation.
• Analyzed filtering performance under abnormal measurements using three Mean Squared Errors (MSEs).
• Developed and evaluated a fault-tolerant CKF with a fading factor for innovation correction.
• Proposed an improved fault detection algorithm.

Main Results:

• The fault-tolerant CKF demonstrates improved performance compared to the standard CKF under abnormal measurements.
• Analysis using MSEs shows the effectiveness of the fault-tolerant approach in maintaining accurate state estimation.
• The proposed fault detection algorithm mitigates sampling approximation deviations inherent in CKF.

Conclusions:

• The fault-tolerant CKF provides a robust solution for state estimation in nonlinear systems facing abnormal measurements.
• Theoretical analysis and numerical examples validate the enhanced performance and reliability of the fault-tolerant CKF.
• The study contributes to the practical application of fault-tolerant filtering techniques in engineering.