Task implementation of synchronous finite state machines

Abstract

Model-based design of embedded control systems using Synchronous Reactive (SR) models is among the best practices for software development in the automotive and aeronautics industry. SR models allow to formally verify the correctness of the design and to automatically generate the implementation code. This improves productivity and, more importantly, can ensure a correct software implementation (preserving the model semantics). Previous research focuses on the concurrent implementation of the dataflow part of SR models, including the optimization of the block-to-task mapping and communication buffer sizing. When the system also consists of blocks implementing finite state machines, as in modern modeling tools like Simulink and SCADE, the task implementation can be further optimized with respect to time and memory. In this paper we analyze problems and opportunities in the implementation of finite state machine subsystems. We define the constraints and efficient policies for the task implementation of such systems.