Abstract
The emerging scenarios of cyber-physical systems (CPS), such as autonomous vehicles, require implementing complex functionality with limited resources, as well as high performances. This paper considers a common setup in which multiple control and non-control tasks share one processor, and proposes a dual-mode strategy. The control task switches between two sampling periods when rejecting (coping with) a disturbance. We create an optimisation framework looking for the switching sampling periods and time instants that maximise the control performance (indexed by settling time) and resource efficiency (indexed by the number of tasks that are schedulable on the processor). The latter objective is enabled with schedulability analysis tailored for the dual-mode model. Experimental results show that (i) given a set of tasks, the proposed strategy improves the control performances whilst retaining schedulability; and (ii) given requirements on the control performances, the proposed strategy is able to schedule more tasks.
| Original language | English |
|---|---|
| Article number | 85 |
| Pages (from-to) | 85:1-85:20 |
| Number of pages | 20 |
| Journal | ACM Transactions on Embedded Computing Systems (TECS) |
| Volume | 18 |
| Issue number | 5s (EMSOFT 19) |
| DOIs | |
| Publication status | Published - Oct 2019 |
| Event | ESWEEK 2019 - Fifteenth Embedded Systems Week: International Conference on Embedded Software Companion (EMSOFT) 2019 - New York, United States Duration: 13 Oct 2019 → 18 Oct 2019 https://ieee-cas.org/conference/esweek-2019-embedded-systems-week |
