Adaptive Distributed Observer-based Model Predictive Control for Multi-agent Formation with Resilience to Communication Link Faults

TL;DR

This paper introduces an adaptive distributed observer-based model predictive control approach that improves multi-agent formation tracking resilience to communication faults by localizing control problems and estimating leader states.

Contribution

It presents a novel adaptive observer-based MPC method that enhances fault tolerance and simplifies control in multi-agent formations with communication issues.

Findings

Improved resilience to communication link faults.

Decentralized control problem decomposition.

Enhanced estimation accuracy of leader states.

Abstract

In order to address the nonlinear multi-agent formation tracking control problem with input constraints and unknown communication faults, a novel adaptive distributed observer-based distributed model predictive control method is developed in this paper. This design employs adaptive distributed observers in local control systems to estimate the leader's state, dynamics, and relative positioning with respect to the leader. Utilizing the estimated data as local references, the original formation tracking control problem can be decomposed into several fully localized tracking control problems, which can be efficiently solved by the local predictive controller. Through the incorporation of adaptive distributed observers, this proposed design not only enhances the resilience of distributed formation tracking against communication faults but also simplifies the distributed model predictive control formulation.