Online Period Selection for Wireless Control Systems
Venkata P. Modekurthy, Abusayeed M. Saifullah · 2019
Recent advancements in Industrial Internet-of-Things and cyber-physical systems through the development of wireless standards like WirelessHART and ISA100 for real-time and reliable communication paved the way for a new industrial trend called Industry 4.0. Industry 4.0 proposes to improve production efficiency by employing smart factories. One approach to improve the production efficiency is to predict the external disturbances and adjust the sampling periods accordingly. In a wireless control system, an online adjustment of the sampling periods can decrease the energy consumption of nodes, especially when the system is within a stable state. Although adjusting the sampling period is beneficial, the stability of the system may be compromised due to external disturbances through an increase in sampling period while decreasing the sampling period can impact the real-time performance and energy of the wireless network. Existing work on online sampling period selection assumes that the controller computes the periods and schedules, and repeatedly broadcasts the new sampling periods and schedules to all the nodes. Such an approach is highly energy consuming and can impact control performance. In contrast, to handle online period selection, we propose an autonomous scheduler based on game theory, where each node in the network generates the schedules locally and without any communication with the others. Our approach can handle any changes in the period and link qualities locally and on the fly. We also propose a heuristic for online sampling period assignment, where each node predicts the state and adjust the period locally. Our evaluation on a case study shows that the proposed approach consumes at least 59% less energy when compared to the state-of-the-art approach.