On Least Idle Slot First Co-scheduling of Update and Control Tasks in Real-Time Sensing and Control Systems

Typical real-time sensing and control systems consist of a set of update tasks for installing sensor measurements from the operation environment and a set of control tasks to access to these measurements for making control decisions. Although configuring the sensors with higher sampling rates could improve the accuracy of the measurements and control quality in general, scheduling high frequent update jobs may seriously affect the schedulability of the control tasks. Missing or delaying the control tasks may severely degrade the overall control performance of the system. In this paper, instead of using the traditional periodic update model, we adopt the a periodic update model in generating update jobs for maintaining data validity. We propose an adaptive co-scheduling algorithm called Least Idle Slot First (LISF) to schedule the update tasks and control tasks with the purposes to meet the deadlines of the control tasks and maximize the quality of control (QoC) offered by the control tasks. LISF schedules the jobs in the ascending order of the number of available idle slots before their deadlines and defers the release times of update jobs as long as the corresponding data objects are maintained within the required quality. The experiment results show that LISF can effectively improve the system schedulability and the control performance in the real-time sensing and control systems.

I. Introduction

With the rapid advancements in computing and communication technologies, the research in Cyber-Physical Systems (CPS) has received tremendous interests in recent years [1],[2]. A typical example of CPS is the wireless sensing and control systems. As shown in Fig. 1, sensors are deployed in the operation environment to monitor the real-time status of various entities. They report sensor measurements through a wireless sensor network [3], [4] to the controller which executes control tasks to generate control decisions to be executed by the actuators. To support the control functions, there are two important types of tasks to be processed at the controller: the update tasks and the control tasks. The update tasks are responsible for installing the sensor measurements into a real-time database (RTDB) [5], and the control tasks are defined according to users' control strategies. The control tasks are invoked periodically and each invocation has a hard deadline on its completion time. Missing the deadline could severely degrade the system performance, i.e., it cannot generate a timely response to a critical event occurred in the operation environment. Thus, in the design of such systems, it is important to guarantee that all the control tasks can be completed before their deadlines. Fig. 1. The architecture of a typical wireless sensing and control system