The Research on a Collaborative Management Model for Multi-Source Heterogeneous Data Based on OPC Communication.

Area of Science:

• Cyber-Physical Systems (CPS)
• Distributed Systems
• Edge Computing
• Industrial IoT

Background:

• Managing multi-source heterogeneous data is a critical challenge in distributed cyber-physical systems (CPS).
• Existing systems struggle with interoperability, state estimation accuracy, temporal asynchrony, and low-latency data distribution.

Purpose of the Study:

• To propose and validate a novel edge-centric computing framework and architecture co-design methodology for ultra-high-performance industrial systems.
• To address challenges in data management, state estimation, sensor asynchrony, and real-time control command delivery.

Main Methods:

• Developed a hybrid OPC communication stack (Client/Server, Publish/Subscribe, P2P) for scalable interoperability.
• Introduced an event-triggered adaptive Kalman filter with online noise-covariance estimation for reduced error and computational load.
• Implemented a Dynamic Time Warping (DTW)-based data-fusion module to resolve temporal asynchrony using Mahalanobis distance optimization.
• Designed a content-aware deterministic message queue with a "rules first" scheduling strategy for guaranteed low-latency control commands (<10 ms).

Main Results:

• State-estimation error reduced by 46.7% and computational load by 41% compared to conventional methods.
• Achieved millisecond-level latency for critical control commands even under heavy data loads.
• Experimental validation with over 100 heterogeneous data sources under industrial-grade workloads.
• Demonstrated simultaneous millimeter-level accuracy in field data acquisition and millisecond-level control command latency.

Conclusions:

• The proposed architecture co-design methodology enables system-level synergy for enhanced performance in industrial CPS.
• The framework provides a promising pathway for developing real-time intelligent systems for next-generation industrial applications.
• The approach is directly applicable to smart manufacturing and other demanding domains requiring high accuracy and low latency.