Related Experiment Videos
Deterministic approach to robust adaptive learning of fuzzy models.
Mohit Kumar1, Regina Stoll, Norbert Stoll
1Institute of Occupational and Social Medicine, Faculty of Medicine, University of Rostock, Germany. mohit.kumar@uni-rostock.de
Summary
This study presents a robust method for adaptive fuzzy system learning, handling data uncertainties and errors without prior knowledge. This approach ensures reliable fuzzy model adaptation for various applications.
Area of Science:
- Intelligent systems
- Control theory
- Fuzzy logic
Background:
- Adaptive fuzzy inference systems are crucial for modeling complex systems.
- Existing methods often require prior knowledge of data uncertainties and modeling errors.
- Handling uncertainties and errors is essential for robust system performance.
Purpose of the Study:
- To develop a robust adaptive learning approach for interpretable Sugeno-type fuzzy inference systems.
- To address data uncertainties and modeling errors without requiring a priori knowledge.
- To enable online learning of membership functions and consequent parameters.
Main Methods:
- Utilizing Hinfinity estimation theory and least squares estimation for parameter updates.
- Implementing a deterministic framework to manage uncertainties and errors.
- Developing a robust approach for online adaptive learning.
Main Results:
- Successfully demonstrated adaptive learning in the presence of data uncertainties and modeling errors.
- Achieved online learning of membership functions and consequent parameters without assumptions on error bounds.
- Validated the robust approach through system identification, time-series prediction, and process estimation.
Conclusions:
- The proposed robust adaptive learning method effectively handles uncertainties and errors in fuzzy systems.
- The approach offers a flexible and reliable solution for online fuzzy model adaptation.
- This work contributes to the advancement of robust and interpretable intelligent systems.