We are studying how the quality of robot movement, perceived robot agency, and blended static/dynamic interactivity between a robot and human performer might influence an audience’s emotional state and belief in the validity of a robot character during a performance.
Robotic animation techniques for live performance typically rely on backstage human puppeteering or playback of pre-rendered animation sequences.
Robotic animation techniques for live performance typically rely on backstage human puppeteering or playback of pre-rendered animation sequences. However, these methods are insufficient for high-speed, close human-robot proximity and coordination, especially when the human performer’s position and timing are unpredictable (ex. rapid passing of objects between human-hands and robot-grippers). Furthermore, simple playback of animation can detract from the believability of the performance if an audience is not convinced that the robot has agency (i.e. its ability to act on its own).
We are developing tools that allow us to compose a human-robot performance that blends pre-rendered choreography with key moments of dynamic interactivity to enhance the realism of the character.
We are developing tools that allow us to compose a human-robot performance that blends pre-rendered choreography with key moments of dynamic interactivity to enhance the realism of the character. If the performance successfully modulates the degree to which the robot responds to the human in a pre-defined manner versus behavior that is completely reactive to the dynamic performer, then the audience might still perceive the robot as having complete agency. For example, as the robot is playing back a choreographed series of poses, it might also track the face of the performer to maintain eye contact. By blurring lines in this interaction, the audience might be more willing to believe the robot is animate.
What is the affect space for a human-magician performance?
• What is the affect space for a human-magician performance?
• Can we improve on current robot animation techniques by including computational choreography and aesthetics-influenced motion planning in ways that lead to desired emotional reactions in observed human-robot collaboration?
• What are challenges and opportunities when designing human-robot performances? • Can we generate a new class of tools and approaches that facilitate artistic and functional robot programming by non-experts?
Publishing robot pose sequences to communities of performers in the same way musicians share scores.
Areas of exploration in choreographic timing of human-robot performances include more dynamic or complex playlist structures (ex. Branching story lines) that take advantage of narrative pacing theory and computationally generation of dramatic timing. We also look forward exploring parametric global adjustment of pose sequences based on narrative arcs and expressive descriptions that abstract away from specific robotic poses (ex. “The energy level should build throughout the performance”). We believe that we can create affective animations or movement “flavors” (ex. Bored, nervous, intimidated) that carry their own impact on timing and pacing, and, in turn, could be triggered by inexpensive sensor data from the human performer; if the magician is moving rapidly towards the robot or speaking at fast pace, for example, the robot might cue more agitated behaviors. Finally, formalizing a syntactical format for human-robot choreography might prove to be useful when sharing and publishing robot pose sequences to communities of performers in the same way musicians share scores.