Functionally and Temporally Correct Simulation for Cyber-Physical Systems

Abstract

When developing a Cyber-Physical System (CPS), simulators are commonly used to predict the final performance of the system at the design phase. However, current simulation tools do not consider timing behaviors of the cyber-system such as varying execution times and task preemptions. Thus, their control performance predictions are far different from the real performance, and this leads to enormous time and cost for a system development, because multiple re-design and re-implementation phases are required, until an acceptable system configuration is determined. Motivated by this limitation, this dissertation proposes functionally and temporally correct simulation for the cyber-side of a CPS. The key idea of the proposed approach is to keep the data and time correctness only at the physical interaction points to maximally enjoy the freedom of scheduling simulated jobs. For this, we transform the simulation problem to a real-time job scheduling problem with precedence constraints necessary for the functional and temporal correctness. Then, we propose an efficient scheduling algorithm for the functionally and temporally correct real-time simulation. The proposed approach significantly improves the real-time simulation capacity of the state-of-the-art simulation methods while keeping the functional and temporal correctness. Our evaluation through both synthetic workload and actual implementation confirms both high accuracy and high efficiency of our approach compared with other state-of-the-art methods.