Unobtrusive Mood Assessment for Training Applications

Many tasks are conducted under time and performance constraints, but the mood-congruency of real situations rarely translate to training environments. The ability to perceive affect (emotion and mood states) is essential in understanding how the trainee is reacting to the scenario. This affective data can be used to modify the challenge and flow of scenario or the amount/frequency of support/direction provided to the trainee. Today, training systems often fail to consider trainee data beyond performance measures and at times miss the opportunity to match mood stimulus in the scenario to learning needs of the trainee. This paper discusses our early results using commodity devices (e.g. Emotiv) to identify mood (a state lasting from several minutes up to several hours) during several mood induction procedures. The mood induction focuses on four strong affective states and uses a variety of methods of induction. Early results are reported.