Adaptive Control of Multiagent Systems for Finding Peaks of Uncertain Static Fields

Abstract

In this paper, we design and analyze a class of multiagent systems that locate peaks of uncertain static fields in a distributed and scalable manner. The scalar field of interest is assumed to be generated by a radial basis function network. Our distributed coordination algorithms for multiagent systems build on techniques from adaptive control. Each agent is driven by swarming and gradient ascent efforts based on its own recursively estimated field via locally collected measurements by itself and its neighboring agents. The convergence properties of the proposed multiagent systems are analyzed. We also propose a sampling scheme to facilitate the convergence. We provide simulation results by applying our proposed algorithms to nonholonomic differentially driven mobile robots. The extensive simulation results match well with the predicted behaviors from the convergence analysis and illustrate the usefulness of the proposed coordination and sampling algorithms.