Nonsmooth Control Barrier Function Design of Continuous Constraints for Network Connectivity Maintenance

TL;DR

This paper introduces a novel method for maintaining network connectivity in multi-robot systems by designing nonsmooth control barrier functions that ensure continuous control inputs while preserving algebraic connectivity.

Contribution

It develops a new approach that leverages nonsmooth analysis and algebraic graph theory to design continuous control barrier functions for connectivity maintenance.

Findings

The proposed method guarantees network connectivity during robot coordination.

Simulations and experiments validate the effectiveness and continuity of the control approach.

The approach successfully handles the nonsmooth nature of algebraic connectivity constraints.

Abstract

This paper considers the problem of maintaining global connectivity of a multi-robot system while executing a desired coordination task. Our approach builds on optimization-based feedback design formulations, where the nominal cost function and constraints encode desirable control objectives for the resulting input. Our solution uses the algebraic connectivity of the multi-robot interconnection topology as a control barrier function and critically embraces its nonsmooth nature. We take advantage of the understanding of how Laplacian eigenvalues behave as their multiplicities change, in combination with the flexibility provided by the concept of control barrier function, to carefully design additional constraints that guarantee the resulting optimization-based controller is continuous and maintains network connectivity. The technical treatment combines elements from set-valued theory, nonsmooth analysis, and algebraic graph theory to imbue the proposed constraints with regularity properties so that they can be smoothly combined with other control constraints. We provide simulations and experimental results illustrating the effectiveness and continuity of the proposed approach in a resource gathering problem.