Rationale

You are driving your car along a familiar route and, all of a sudden, you feel a sensation, as if someone is carving an S on your back. The context you are in suggests something is affecting your trip. Right after, the same S-shape is played back through the car audio system, and your car mate perceives it as well. A few seconds later both of you see a choreographed set of road safety lights indicating how the normal route is being deviated, due to an accident or to roadwork. You drive more carefully, and get back home with just a small delay.
The evoked scenario shows how short-term projections into the future could be obtained through multisensory and multiscale trajectories.
Trajectories are everywhere, at all scales, from inside our body to outer space: shivers along the spine, mosquitos around our head, mexican waves at the stadium, dynamic road curve warning systems, bird flocks, comets.
Trajectories carry meaning through their attributes, from the basic to the complex: direction, speed, curvature, thickness, color, shape, texture, timbre, which can be expressed in bases of primitives. For example, an utterance can be expressed as a sound trajectory in a space of primitives such as phonation, turbulence, and pulses.
Human gestures are trajectories. Through musical gestures, dance moves, javelin throws, handwritten strokes, caresses, humans produce trajectories that make an act manifest and expressive to other humans. Modeling trajectories at different spatio-temporal scales and across sensory modalities would enable the consistent design of gestural interfaces, from the intimate to the public space.
Traces are spatial leftovers of trajectories, a sign that something has happened or existed, that can possibly be followed. They can be more or less ephemeral, remain on a physical support or just as mental imagery, as a snake track on sand, or a music glissando in auditory memory. The geometry of trajectories and of related traces can be manipulated through interaction of the senses. For example, a harsh visual trajectory can be perceived and remembered as smoother if displayed together with a round sound. Trajectories are expressive and social. Classic studies, as well as contemporary follow-ups, have been showing that very simple moving objects, extremely sparse in space and time, are sufficient to characterize trajectories with higher-level qualities, such as intentions, emotions, and causal relationships. In social interaction, trajectories that produce coordinated movement are a key aspect of joint action, e.g., in team sports and ensemble music or dance performances. Entrainment describes the temporal dynamics of interacting rhythmic systems. Movement features of trajectories are relevant for perceiving and establishing entrainment and synchronization between cooperating humans.
Trajectories and traces are the constituents of sketching, an ubiquitous, upfront, and indispensable activity of all human design practices. There is a strong urge to develop sketching tools that are inherently multisensory, so that immediate human actions can be reliably converted into sketches that can be communicated across different sensory modalities. We can sketch with a pencil, with voice, or with full body, and technology should seamlessly help us make multisensory sketches, through trajectory analysis and synthesis.
Computational approaches have already been successfully applied to extract expressive and social information from trajectories but, in most of the cases, they either address a particular aspect or deal with a specific sensory modality. Moreover, analysis is usually performed on a single temporal scale, suitable for the task at hand. A comprehensive investigation of trajectories with particular reference to models and computational approaches leveraging supramodal representations over multiple temporal and spatial scales is still missing and highly needed.
The rationale of the MAHATMA project is that it is possible to derive models, in the form of mathematical equations or computational structures, of trajectories of human expression, and emulate them by artificially generated trajectories. These models should be adaptable to different levels of proximity and be also supramodal, thus encoding features of trajectories that can be cast into sense-specific renderings. Tools can be developed that convert and extend trajectories that are the result of human motion and coordination, across different sensory domains and at different scales. Such tools can be applied to improve performance and the effectiveness of training in sports sciences, and they can be exploited to sketch and interactively design trajectories for multisensory displays and the performing arts.