A Novel Approach to Implement Takagi-Sugeno Fuzzy Models

Area of Science:

• Control Systems Engineering
• Computational Intelligence
• Nonlinear System Modeling

Background:

• Complex nonlinear systems require advanced modeling techniques.
• Existing Takagi-Sugeno (T-S) fuzzy models have limitations in flexibility and computational load.
• Accurate modeling is crucial for effective control system design.

Purpose of the Study:

• To propose new fuzzy c-regressing model algorithms for T-S fuzzy modeling.
• To develop methods for modeling nonlinear systems with polynomial or state-space consequents.
• To extend fuzzy modeling capabilities for scenarios with limited or unavailable input-output data.

Main Methods:

• Development of the fuzzy c-regression state model (FCRSM) algorithm.
• Introduction of FCRSM-ND for modeling with available controllers but no precollected data.
• Design of FCRSM-FREE for simultaneous online tuning of fuzzy controllers and system models.
• Utilizing input-output data for antecedent and consequent identification.

Main Results:

• The FCRSM algorithm demonstrates a low computation load by considering a single input variable in the antecedent.
• The proposed models can represent unknown systems in both polynomial and state-space forms.
• FCRSM-ND and FCRSM-FREE algorithms effectively address data scarcity and online tuning requirements.
• Numerical simulations confirm the efficacy of the developed fuzzy modeling approaches.

Conclusions:

• The proposed fuzzy c-regressing model algorithms provide a flexible and efficient approach to T-S fuzzy modeling of nonlinear systems.
• The FCRSM framework, including its extensions FCRSM-ND and FCRSM-FREE, enhances modeling capabilities for various practical scenarios.
• These algorithms offer significant advantages in terms of computational efficiency and modeling versatility for complex systems.