Joint Task Scheduling and Communication-Computation Optimization for Wireless Networked Control with HRLLC

This paper studies a wireless control system at network edge, in which a base station (BS) wirelessly coordinates the closed-loop control of multiple subsystems each consisting of a plant, a sensor, and an actuator. In this system, the BS first collects the state information from the sensors of plants, then processes the information via edge computing, and finally sends the obtained command signals back to the actuators for controlling the plants. In particular, we consider the hyper-reliable and low-latency communications (HRLLC) for the state and command signal transmission, by using the rate formulas based on short-packet communication. Under this setup, we first present a time-division-multiple-access (TDMA) protocol for coordinating the sensing, communication, and computation among the multiple plants. Then, we jointly optimize the task scheduling as well as the communication and computation resource allocations to minimize the closed-loop control latency while ensuring the stability of the multiple control subsystems. The considered problem is a highly non-convex combinatorial optimization problem that is difficult to solve. To resolve this issue, we present efficient algorithms by first optimizing the communication and computation resource allocations under given task scheduling via the techniques of alternating optimization and successive convex approximation, and then designing the task scheduling based on the exhaustive search or the low-complexity flow-shop scheduling. Numerical results show that the proposed joint resource allocation design with exhaustive search based task scheduling significantly outperforms other benchmark schemes without such joint optimization, and the proposed low-complexity task scheduling based on flow-shop scheduling achieves performance close to the upper bound by exhaustive search.