Communication through motion in dance with topological constraints

Abstract

This paper summarizes work that investigates human interactions in terms of communication through motion in cooperative control. We are interested in multi-agent distributed control models with a shared goal and with the agents being required to accomplish secondary objectives. The study presented here extends the earlier work on the analysis of leader-follower interactions in salsa. The move transitions in salsa are described by simple rules suggested by topological knot theory. The initial and final poses of each move in the dance are represented by link diagrams. The link invariant Alexander Polynomial is calculated for classified link diagrams so that the physical motion primitives in dance are described as polynomial function manipulations. The size of the alphabet for the dance moves given in earlier work is extended by introducing three new moves. The leaders' decision process in a transition system is discussed and it is compared with the prior models which had fewer moves. Complexity metrics are proposed to capture the new alphabet and the proficiency hierarchy in the dance.