A new method for control of nonlinear networked systems

• The actual plant input is known to the controller at each time step, therefore, it can cope with relatively large network delays and packet losses.
• A new controller design methodology is proposed for nonlinear systems, by taking the delay variation into account.
• Sampling period is adapted based on the network condition.
• The tracking issue is addressed well.

Networked control of a class of nonlinear systems is considered. For this purpose, the previously proposed variable selective control (VSC) methodology is extended to the nonlinear systems. This extension is based upon the decomposition of the nonlinear system to a set fuzzy-blended locally linearized subsystems, and further application of the VSC methodology to each subsystem. Using the idea of parallel distributed compensation (PDC) method, the closed-loop stability of the overall networked system is guaranteed, using new linear matrix inequalities (LMIs). For the real-time implementation, real-time control signals are constructed for every entry of pre-specified vector of time delays, which is selected based on the presumed upper-bound of the network time delay. Similar to the traditional packet-base control methodology, such control signals are then packed as a control-side packet and transmitted back to a time delay compensator (TDC) located on the plant-side of the network. According to the most recent network time delay, the TDC selects just one entry of the control vector and applies it to the actuator through a zero order hold element. A sufficient condition for closed-loop asymptotic stability is determined. Simulation studies on nonlinear benchmark problems demonstrate the effectiveness of the proposed method.

Figure optionsDownload as PowerPoint slide