Optimized Distributed Filtering Over Binary Sensor Network: A Dynamic Event-Triggering Protocol With Token Bucket...

Area of Science:

• Control Engineering
• Signal Processing
• Systems Science

Background:

• Distributed filtering is crucial for systems with decentralized sensors and limited communication.
• Binary measurements introduce uncertainties that challenge traditional filtering approaches.
• Event-triggered communication protocols are needed to manage network resources efficiently.

Purpose of the Study:

• To develop an optimized distributed filtering strategy for discrete linear time-varying systems using binary measurements.
• To address uncertainties in binary measurements with a novel time-varying threshold.
• To implement dynamic event-triggered protocols for resource-constrained information transmission.

Main Methods:

• Designing two cases for extracting measurement information from binary data.
• Introducing a time-varying threshold strategy to mitigate measurement uncertainties.
• Employing dynamic event-triggering protocols with token bucket specifications for data transmission scheduling.

Main Results:

• Ensuring exponential boundedness in the mean square for the filtering error system.
• Recursively calculating filter parameters via distributed optimization problems.
• Demonstrating the effectiveness of the proposed distributed filtering scheme through simulations.

Conclusions:

• The developed distributed filtering scheme effectively handles binary measurements and system uncertainties.
• The combination of novel thresholding and event-triggering enhances filtering performance and scalability.
• The approach provides a robust solution for optimized distributed filtering in networked systems.