Plan and Goal Recognition System for Adaptive Pilot Assistance in Tactical Helicopter Operations

In our contribution, we present an approach to an adaptive assistant system that supports pilots of a future military helicopter in dynamic missions by determining their intents on a tactical level. In a highly dynamic military mission environment, a pre-defined mission plan can quickly deteriorate due to unpredicted situational changes. To assist proactively and cooperatively, the assistant system needs an understanding of the currently pursued task and the corresponding course of action. This article describes an approach for a pilot assistant system that adapts its support to the momentary tactical plan to assist in the mission execution by observing the pilot interactions and probabilistic reasoning. For that, a plan and goal recognition system, based on hierarchical task networks, is integrated to generate partial-order sequences of observable actions. To derive the specifications of the observation and assistance demand a single-subject case study was conducted. To provide adaptive assistance the functions for an online intent recognition need to be developed and integrated into our mission and cockpit simulation environment.