Event-Triggered Recursive Filtering for Shift-Varying Linear Repetitive Processes.

Area of Science:

• Control Systems Engineering
• Signal Processing
• Networked Systems

Background:

• Recursive filtering is crucial for estimating system states.
• Limited network resources pose challenges for real-time data transmission.
• Shift-varying linear repetitive processes (LRPs) require specialized filtering techniques.

Purpose of the Study:

• To design an event-triggered recursive filter for shift-varying LRPs.
• To ensure and optimize the filtering error variance under resource constraints.
• To reduce network load by transmitting data only when necessary.

Main Methods:

• Developed a novel event-triggered communication strategy.
• Transformed LRPs into a general Fornasini-Marchesini model using lifting technique.
• Introduced a new triggering-shift sequence and event-triggered rule.
• Utilized mathematical induction to guarantee an upper bound on filtering error variance.
• Solved Riccati-like difference equations to optimize filter gains.

Main Results:

• Successfully designed a recursive filter operating under an event-triggered mechanism.
• Ensured an optimized upper bound on the filtering error variance.
• Demonstrated the monotonicity of filtering performance with respect to the event-triggering threshold.
• Validated the proposed filtering scheme through simulations.

Conclusions:

• The event-triggered recursive filtering approach is feasible and effective for shift-varying LRPs.
• The proposed method significantly reduces network resource occupancy.
• Optimized filter gains lead to improved filtering performance and error bounds.