The design and simulation of networked control systems with online extreme learning machine PID

Network-based control systems (NCSs) for industrial processes recently have been paid much more attention. However, the control engineers must face to an unavoidable problem of whether or not the traditional proportional-integral-derivative (PID) controller for the network case is available. More specifically, how to design PID controller in the presence of the networkinduced time-varying delay? In this paper, an online extreme learning machine-based PID controller is proposed, where the Jacobian information relating to online extreme learning machine is identified, and then is used to adjust the PID parameters. Lastly, a simulation using TrueTime blocks in a Simulink/MATLAB box is given, and it is observed that the closed loop system has good dynamic and adaptive performance.