Motive and Affect-based Robotic Control (MARC)

This project addresses two core challenges to wider and more effective use of robotic and uninhabited vehicles (UxVs), the need for more robust and understandable autonomy, and for simplified, intuitive, and more learnable human control mechanisms. The Phase I effort created an advanced control mechanism and interface for UxVs based on PAC, an innovative cognitive model of personality and emotion. The resulting PAC-based design is named MARC -- Motive and Affect-based Robotic Control. MARC is intended to translate between ‘ordinary personality and emotion language’ – the kind the human UxV controller might use in communicating with other people -- and underlying computational structures that control the UxV. Phase I used the biologically inspired PAC software to build a proof-of-concept MARC prototype that controlled multiple simulated UxVs. A human-in-the-loop experiment compared the MARC prototype with a more conventional UxV control mechanism, and found that MARC led to significantly fewer errors, lower workload, and higher mission accomplishment in comparison to the conventional control paradigm. The goal of the Phase II effort (base and option) is to bring MARC to TRL 6 through complete component development and demonstration of an integrated MARC in a relevant operational environment.