Hierarchical non-intrusive in-situ requirements monitoring for embedded systems

Accounting for all operating conditions of a system at the design stage is typically infeasible for complex systems. In-situ runtime monitoring and verification can enable a system to introspectively ensure the system is operating correctly in the presence of dynamic environment, to rapidly detect failures, and to provide detailed execution traces to find the root cause thereof. In this paper, we seek to address two challenges faced in using in-situ runtime verification for embedded systems, including (1) efficiently defining and automatically constructing a requirements model for embedded system software and (2) minimizing the runtime overhead of observing and verifying the runtime execution adheres to the requirements model. We present a methodology to construct a hierarchical runtime monitoring graph from system requirements specified using multiple UML sequence diagrams, which are already commonly used in software development. We further present the design of on-chip hardware that nonintrusively monitors the system at runtime to ensure the execution matches the requirements model. We evaluate the proposed methodology using a case study of a fail-safe autonomous vehicle subsystem and analyze the relationship between event coverage, detection rate, and hardware requirements.