On-Demand Transmission for Edge-Assisted Remote Control in Industrial Network Systems

• Published in: IEEE transactions on industrial informatics Vol. 16; no. 7; pp. 4842 - 4854
• Main Authors: Chen, Cailian, Lyu, Ling, Zhu, Shanying, Guan, Xinping
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.07.2020; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Communication networks; Computer architecture; Control centres; Control systems design; Detection; Edge-assisted remote control; Energy budget; Estimation; estimation convergence; Feedback control; industrial network system; Network control; Nonlinear programming; on-demand transmission; Parameter estimation; Process parameters; Reliability; Remote control; Remote sensing; Revenue; Sensors; Systems stability; Wireless communication; Wireless sensor networks
• ISSN: 1551-3203; 1941-0050
• DOI: 10.1109/TII.2019.2951472

Sensing data and control commands are frequently exchanged over communication networks for remote data acquisition and distributed control in industrial network control systems. The system performance relies on the design of sensing, transmission, and control. Due to the harsh environment in industrial field and the limited network resources, it is very challenging to meet the high requirement on transmission reliability for remote feedback control. In order to enhance the transmission ability for this kind of systems, an edge-assisted system architecture is proposed for the sensing and control processes. The parameter estimation for the sensing process is executed in the so-called edge estimator, and the controller is designed in a remote control center. Under this architecture, in this article an on-demand transmission scheme is designed by characterizing the overall effects of transmission reliability on the estimation and control performance. The overall system is optimized by formulating a revenue-cost maximization problem subject to the constraints of system stability, estimation convergence, spectrum utilization, and energy budget. The formulated mixed-integer nonlinear programming problem can be effectively solved with the block coordinate descent method. It is decomposed into two subproblems over disjoint variable sets. Simulation results demonstrate the advantages on both network-wide revenue and control-transmission cost.