Constrained Control for Optimal Collaborative Applications

Shared human-robot workspaces are becoming a prominent part of the modern industrial paradigm. These collaborative environments often present highly dynamic features. It is thus fundamental that the robot maintains a behavior which is both robustly stable during the interaction with the environment and both flexible enough to ensure safety and reactivity to the varying conditions of the workspace.

In this paper, we propose a control architecture which satisfies both these conditions by combining an energy tank based variable admittance framework with control barrier functions.