Static timing analysis of cyber-physical systems with relaxed real-time constraints

Overview & Synopsis

Modern Cyber-Physical Systems (CPS) often involve dependent real-time processes bound to different timing constraints. These CPSs can be modeled using PolyGraph, a formalism that extends the Cyclo-Static Dataflow model with explicit real-time constraints. Unlike traditional real-time approaches where timing constraints are specified for every process, PolyGraph can offer greater flexibility by binding real-time constraints for only a subset of processes (e.g., processes associated with hardware components such as sensors and actuators), along with end-to-end latency specifications. In that case, processes without specified timing constraints only inherit ones from processes with explicit and specified timing constraints due to data dependencies. This paper aims to formalize a method to derive timing constraints for an entire PolyGraph model where only a subset of processes has explicit timing constraints in the model. We demonstrate how this approach eases scheduling constraints and enables early detection of missed deadlines through two examples from a vehicle ADAS and the Ingenuity Mars helicopter.