Augmenting Human Dance Performance With a Robot Arm: An Incremental Dynamic Mapping Method

In artistic applications of robots as dancing bodies, dimensionality reduction techniques like principal component analysis (PCA) are commonly used to project complex human motion into low-dimensional embedding spaces, making it suitable for robots with fewer degrees of freedom. However, standard PCA requires access to complete motion trajectories and fails to account for temporal correlations, thereby limiting real-time adaptability and restricting the dancer’s improvisational freedom.