We have developed a cooperative simulation environment for multiple electronic control units (ECUs) including a parallel executions mechanism to improve the test efficiency of a system, which was designed with multiple ECUs for autonomous driving. And we have applied it to a power window system for multiple ECUs with a controller area network (CAN). The power window model consists of an electronic-mechanical model and a CPU model. Each simulator with a different executions speed operates in parallel using a synchronization mechanism that exchanges data outputted from each simulator at a constant cycle. A virtual ECU simulated microcontroller hardware operations and executed its control program step-by-step in binary code to test software for the product version. As co-simulation technology, a mechanism that synchronously executes heterogeneous simulators and a model of an in-vehicle communication CAN connecting each ECU were developed. We have evaluated the simulation speed using multiple virtual power window systems with parallel executions. The simulation speed is 60% faster using parallelization. This method was applied to the ECU model and mechanical model system that was found to work together in a virtual prototyping system. This virtual prototyping system can synchronously execute simulators running on different PCs. This means that a large-scale simulation including a detailed simulation of binary code can be performed by executing a large number of simulators in parallel.
Evaluation of Parallel Executions on Multiple Virtual ECU Systems
Sae Technical Papers
WCX World Congress Experience ; 2018
03.04.2018
Aufsatz (Konferenz)
Englisch
Evaluation of Parallel Executions on Multiple Virtual ECU Systems
British Library Conference Proceedings | 2018
|Detecting data races in parallel program executions
TIBKAT | 1989
|Visualizing RCE Workflow Executions via W3C Provenance
IEEE | 2023
|