During a deliberate attack on an insurgent-held city, a Marine infantry company receives fire from a small building next to a mosque. What should the artillery Forward Observer (FO) do? The answer depends on context. If the fire coming from the building causes casualties, the FO should conduct an Immediate Suppression mission. If the insurgents' fires do not have any effects on the Marines below and they can take cover, the FO needs to formulate a course of action with the company commander.

How would we measure FO performance in simulator-based training for this scenario? It's not enough simply to take obvious measurements like target location error or target/ammunition combination. We must have an understanding of the FO's context, and measure and assess the FO's performance accordingly. The performance measurement infrastructure in the training environment must support these activities.

In this talk, we discuss a formal representation of context for human performance measurement in immersive training environments and how that representation fits into an innovative language for expressing those measurements, Human Performance Measurement Language (HPML). We show how context plays a role both as triggers for measurements and as key information for assessments, and demonstrate a method for convenient elicitation of context information from expert instructor/operators. We provide illustrative examples of training mission contexts and show how they may be represented using this formalism.

We further discuss how we have implemented context representation capabilities in real-world simulator-based training situations, including a forward observer simulator operating in a variant of a Distributed Virtual Training Environment-based federation and an F/A-18 simulator flying in a Navy Aviation Simulation Master Plan-based federation. We conclude by discussing additional benefits of representing context in simulator-based training environments.