An active disturbance rejection control approach for decentralized tracking in interconnected systems

In this article, the local trajectory tracking control problems -on the part of agents acting on a complex interconnection of nonlinear subsystems- are reformulated as active disturbance rejection control (ADRC) problems. The new approach gives rise to a more effective and practical robust decentralized solution, with virtually no information on local plants interaction dynamics and limited knowledge of each subsystem model. The key idea is that each agent regards all the nonlinear uncertainties arising from: 1) unknown state dependent nonlinearities of his own subsystem, 2) the environmental exogenous effects and 3) the rest of interaction effects coming from the network interconnection structure, as a lumped, unstructured, time-varying disturbance that perturbs the agent´s subsystem local dynamics. It is shown that such disturbance can be actively estimated and canceled from each individual subsystem model dynamics, thus facilitating the accomplishment of the individual control objectives in a direct fashion. An illustrative example is presented dealing with two dynamically interconnected set of gantry cranes, subject to uncertainty and interaction while trying to independently track a prescribed position reference trajectory in a decentralized manner.