MDLn: A Motion Description Language for Networked Systems

Abstract

In this paper we extend the definition of a Motion Description Language (MDL) to networked systems. This new construction (MDLn) supports inter-agent specification rules as well as desired network topologies, enabling us to specify high-level control programs for group interactions. In particular, MDLn-strings specify multi-modal executions of the system through a concatenation of modes. Each mode in the MDLn-string is a triple, specifying a control law, interrupt conditions, and desired network dependencies. In addition to proposing MDLn as a specification language for networked systems, we also give an architecture in which MDLn strings can be effectively parsed and executed in multi-robot applications.