Adaptive event-triggered-based efficient model predictive control for nonlinear systems subject to cyber attacks and...

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Networked Systems

Background:

• Nonlinear systems are challenging to control due to complexities like cyber attacks and actuator saturation.
• Existing model predictive control (MPC) methods can be communication-intensive and computationally demanding.

Purpose of the Study:

• To develop a novel event-triggered-based decay aggregation efficient model predictive control (DAEMPC) for nonlinear interval type-2 T-S fuzzy systems.
• To address cyber attacks (e.g., denial-of-service) and actuator saturation in networked control systems (NCS).
• To improve communication resource utilization and reduce computational burden.

Main Methods:

• An adaptive event-triggered (AET) strategy was employed to optimize data transmission between sensors and controllers.
• A Bernoulli random process modeled denial-of-service (DoS) attacks, while polytopic description handled actuator saturation.
• The DAEMPC involved offline computation of control laws and feasible sets, with online optimization of perturbation variables.

Main Results:

• The AET strategy effectively managed communication resources in the NCS.
• The DAEMPC algorithm successfully compensated for network deficiencies and enlarged the initial feasible set.
• The proposed method demonstrated a significant reduction in computational burden compared to traditional approaches.

Conclusions:

• The AET-based DAEMPC is a valid and effective control strategy for nonlinear NCS.
• This approach offers improved performance in terms of communication efficiency and computational load.
• The method's validity was confirmed through simulations on a continuous stirred tank reactor (CSTR).