Network-constrained adaptive control with a nonlinear tracking function

We consider a problem where several agents need to achieve a common control objective. The motivation for this problem is control of power generation in a wind farm under weak-grid conditions, with the individual turbine-generators acting as agents. The control objective is a nonlinear function of the state of the agents. Furthermore, this nonlinear function depends on the parameters of the network connecting the agents, and these parameters are unknown. Based on a newly developed nonlinear adaptive control methodology, a process is designed that simultaneously learns the network parameters and, using individual controllers, drives the agents to their respective steady-state solutions. We also compare the performance of this design to an empirical control design that requires less communication.