Affect-Influenced Control of Unmanned Vehicle Systems

We propose applying recent research in cognitive models of affective behavior to create a system able to modulate the actions of unmanned vehicles in both gross and subtle ways. Using the new interface, instead of simply specifying controller parameters or cost functions, the operators will adjust affective parameters, which will influence the performance of the vehicle(s) over the life of the mission. These intuitive, high-level affective settings such as aggressiveness, stealthiness, and tenacity will allow the operator to set the mission characteristics at a very abstract level. A major thrust of this project will be to develop approaches and techniques to convey a clear and valid understanding to the operator and embody them in the form of a transparent graphical user interface and performance simulator. The software architecture will consist of two major systems, an operator interface component and an embedded control component which can be added to existing unmanned vehicle autonomy system. Both long- and short-term time-varying affective phenomena, such as traits, moods, attitudes, and emotions, will allow the system to provide performance consistent with varying situations. An initial prototype will be developed during Phase I to allow evaluating the overall feasibility of the approach.