Event-triggered state estimation for networked systems with correlated noises and packet losses

Area of Science:

• Control Systems
• Signal Processing
• Information Theory

Background:

• State estimation is crucial for system monitoring and control.
• Traditional methods often involve continuous data transmission, leading to high bandwidth usage.
• Correlated noises and packet losses pose significant challenges in real-world systems.

Purpose of the Study:

• To develop an event-triggered state estimation approach for systems with correlated noises and packet losses.
• To reduce data transmission rates while maintaining acceptable estimation performance.
• To analyze the trade-off between transmission rate and estimator accuracy.

Main Methods:

• A novel event-triggered communication mechanism is proposed, sending data only when a specific condition is met.
• A new event-triggered state estimator is designed, incorporating trigger threshold and correlation coefficient.
• The performance of the estimator is evaluated, and boundedness conditions for covariance expectation are derived.

Main Results:

• The event-triggered approach effectively reduces data transmissions.
• The proposed estimator mitigates the impact of correlated noises and packet losses.
• Adjusting the trigger threshold allows for a tunable trade-off between transmission rate and estimation performance.
• Theoretical guarantees on estimator performance (covariance expectation boundedness) are established.

Conclusions:

• The developed event-triggered state estimation method offers an efficient solution for networked systems.
• The approach provides a practical way to balance communication load and estimation accuracy.
• The findings are validated through application to a target tracking system.