Model predictive control of non-linear systems over networks with data quantization and packet loss

Area of Science:

• Control Systems Engineering
• Networked Control Systems
• Non-linear System Analysis

Background:

• Networked control systems (NCS) introduce challenges like data quantization and packet loss.
• Non-linear systems require robust control strategies to maintain stability.
• Tagaki-Sugeno (T-S) models are effective for representing non-linear plants.

Purpose of the Study:

• To develop a model predictive control (MPC) strategy for non-linear NCS with quantization and packet loss.
• To ensure the stability of the closed-loop system under uncertain communication conditions.
• To provide a systematic method for designing a fuzzy predictive controller.

Main Methods:

• Representing the non-linear plant using a Tagaki-Sugeno (T-S) model.
• Applying the sector bound approach to model data quantization as uncertainties.
• Modeling packet loss using a Bernoulli process.
• Designing a fuzzy predictive controller by solving linear matrix inequalities (LMIs).

Main Results:

• A fuzzy predictive controller was designed to guarantee closed-loop stability.
• The controller effectively handles data quantization and packet loss in the NCS.
• The proposed method was validated through a numerical example, demonstrating its effectiveness.

Conclusions:

• The developed MPC approach provides a stable control solution for non-linear NCS with communication imperfections.
• The use of LMIs offers a computationally feasible method for controller synthesis.
• This research contributes to robust control design in networked environments.