An efficient identification approach for stable and unstable nonlinear systems using Colliding Bodies Optimization...

Area of Science:

• Control Systems Engineering
• Computational Intelligence
• Nonlinear System Identification

Background:

• Hammerstein models are crucial for representing nonlinear dynamic systems.
• Accurate identification of both stable and unstable nonlinear processes is essential for effective control.
• Existing stochastic algorithms face challenges in precision and robustness.

Purpose of the Study:

• To develop an efficient evolutionary algorithm-based approach for identifying Hammerstein models.
• To evaluate the performance of the Colliding Bodies Optimization (CBO) algorithm in system identification.
• To demonstrate the robustness and practical applicability of the CBO algorithm.

Main Methods:

• Utilizing the Colliding Bodies Optimization (CBO) algorithm, an evolutionary computation technique.
• Applying the CBO algorithm to identify stable and unstable nonlinear Hammerstein models and their closed-loop counterparts.
• Assessing performance using output Mean Squared Error (MSE), precision, and accuracy metrics.

Main Results:

• The CBO algorithm achieved minimal output MSE, indicating low bias and variance.
• Consistent and close estimation of output parameters, even in the presence of outliers.
• Superior performance in terms of MSE, computational time, and statistical properties compared to existing stochastic algorithms.

Conclusions:

• The CBO algorithm offers an efficient and robust solution for Hammerstein model identification.
• The proposed method demonstrates practical usefulness and general applicability for system identification problems.
• CBO-based identification schemes provide enhanced accuracy and efficiency over traditional approaches.