Timing analysis for UAS application software

Unmanned Aerial Systems (UAS) are quickly becoming common place, with many of these systems being used to perform tasks autonomously. Such systems will need to be configurable to support a variety of missions, and may require flight software reconfiguration on a per-mission basis. New rapid verification methods are needed to support the timely verification of this software. Such software will likely be built in a component-based manner using a suitable middleware to facilitate rapid development. These systems will likely also rely on an autonomy engine due to the nature of their missions. In this paper we consider one particular middleware (the cFE/cFS) and autonomy engine (PLEXIL) and introduce a model-based approach for end-to-end, stimulus-to-response timing analysis of flight software systems built using them. The method uses a Colored Petri Net based timing model to analyze the behavior of (1) the cFE/cFS framework, (2) the PLEXIL executive, (3) the typical applications that run within cFE/cFS, and (4) the environment of the system. The model is constructed in a generic, parameterized fashion which allows for automated generation of new system configurations from the architectural models of the system. The ability to consider various system configurations without model reconstruction indicates the suitability of the approach to rapid system verification goals. Application of the analysis method requires knowledge of (1) the autonomy plan executed by PLEXIL, (2) the behavior and worst-case response times of the cFS applications, and (3) the environmental stimuli under consideration.