Person - to - Person Contact (P2P)

The paradigm for person-to-person interaction is dance or sports training, where the trainer, for example, teaches a learner through direct limb-to-limb contact. A variant of this scenario is where one of the people is replaced by an entirely programmed virtual character. There are several fundamental research issues to consider. In this proposal presence is more than just the ‘feeling of being there’. It is concerned with how people respond to events. They are present if they respond to events as if they were real at every level of response from unconscious physiological to high level cognitive responses. Of course physical events in the real world are by definition presence inducing. The question only becomes interesting for virtual events within a mixed reality. If we take ‘dancing’ as a paradigmatic example of this scenario then there is presence to the extent that the dancers respond to one another as if they were really dancing, or if one of the partners is virtual that the real human partner responds to the virtual dancer as if real. The highest level question then is the extent to which a virtual representation of a person can substitute for a real person, in this specific domain.

Continuing with our paradigmatic example, a dance has a number of phases: an introduction, a handshake, the dance itself, and finally the termination. Then there are several major research issues in this scenario. The first is the establishment of ground truth – what actually happens that is recordable when two people physically meet and dance? This involves gathering data at many levels – the forces involved in the limb-tolimb interaction, the physiological, behavioral, motor, and subjective responses of the participants.

Second is construction of the model of the dance. In the mixed reality representation, each person would always be dancing with a ‘model’ of the partner, not with the partner directly. There are then two extremes in this situation: the model is wholly computer generated or the model is driven by a completely tracked person at the remote site. The true situation would always be a mix between these extremes. The third problem is the construction of the whole as a set of building blocks that make it possible to explore substitution of different elements for each other. For example, what happens when the visual representation is improved at the cost of the haptics, or vice versa? As we move from the situation of the real person driving the model through tracking towards the model being driven entirely by a programmed agent, at what point does the real human partner notice the difference and presence break down? Another interesting aspect of this building block approach is how the whole dance system can be incorporated as one unit within a higher level application, such as its utilization in a psychotherapy program for social phobia? A fourth major problem is simply the integration of many different elements. This should not be thought of as mere ‘implementation’ since it is exceedingly difficult to synchronize the graphical, auditory and haptic elements, especially when data is moving across a communication network with uncertain performance.