Dependability in robotics as a consensus problem

This paper identifies dependability concerns with concepts well established in systems theory, namely in the field of consensus problems. Implementations in areas characterized by the presence of important uncertainties, such as social robotics, can often be modeled as networks of complex components. Each node in such networks can be given a specific structure including performance indicators which are required to reach consensus. Identifying dependability with the consensus of a set of indicators enables the use of known results of the latter. The paper describes simulation examples on robot formations to illustrate the ideas discussed.