Fuzzy Computing Projects
Fuzzy Filtering Algorithms
To deal with the real-world problems characterized by complexities and uncertainties, the design of the fuzzy filters is an important issue since the real-world applications require the filtering of uncertainties from the experimental data. Thus, we studied the adaptive fuzzy filtering problem based on different estimation criteria. The studied mathematical criteria include followings:
# Robust Regularized Least-Squares Estimation
# H-infinity Optimal Estimation
# Least-Squares Estimation
# Generalized Least-mean-squares like p-norm Algorithms
# Risk-sensitive Estimation
A mathematical theory was developed for the stability, robustness, and steady-state analyses of fuzzy filtering algorithms. This facilitates a computational paradigm to implement an intelligent behavior in the sense of handling uncertainties related to the modeling of the complex and uncertain processes.
Fuzzy Evaluation of Heart Rate Signals for Mental Stress Assessment
This work, being supported by Center for Life Science Automation and Institute for Preventive Medicine of Rostock University, presented a novel method of heart rate signal analysis for stress assessment using fuzzy clustering and robust identification techniques. The emphasis of this study was to handle the uncertainties, arising due to a difference in the physiological behavior of individuals (because of different body conditions, age, gender, and so on), using a fuzzy model.
Fuzzy Physical Fitness Estimation
The aim of this project supported by European Space Agency was to develop a fuzzy expert system for estimating the physical fitness of a patient on a scale from 0 to 1. The complex and uncertain relationships between some of the physiological parameters measurements and opinion of a medical expert about the fitness level of the patient are modeled using a fuzzy inference system.
Computational Intelligence in Drug Design
The Center for Life Science Automation Rostock supported a project for carrying out a research on computational intelligence based drug design. The aim of the work was to develop a fuzzy based computer model for a prediction of the drug activity of a chemical compound from its structure. The computational intelligence techniques were used to remove some of the bottlenecks in structure-activity modelling poroblems.
Fuzzy Systems in Modeling the Environmental Behavior of Chemicals
A joint work with the Institute of Chemistry of Rostock University involved the modeling of the environmental behavior of chemicals. A fuzzy based expert system was developed for predicting the toxicity and bioconcentration factor of chemicals.
Fuzzy Filtering for Physiological Signal Analysis
This study suggests the use of fuzzy filtering algorithms to deal with the uncertainties associated to the interpretation of analysis of physiological signals. The signal characteristics, for a given situation or physiological state, vary for an individual over time and also vary among the individuals with the same state. These random variations are due to the several time-varying factors related to the physiological behavior of individuals which can't be taken into account in the interpretation of signal characteristics for solving a medical decision making problem. The approach is to reduce the effect of random variations on the analysis of signal characteristics via filtering out randomness or uncertainty from the signal using a nonlinear fuzzy filter.
Variational Bayes for a Mixed Stochastic/Deterministic Fuzzy Filter
This study, under variational Bayes (VB) framework, infers the parameters of a Takagi-Sugeno fuzzy filter having deterministic antecedents and stochastic consequents. The motivation of the study is to take advantages of the VB framework in designing fuzzy filtering algorithms. These advantages include an automated regularization, incorporation of statistical noise models, and model comparison capability. The VB method can be easily applied to the linear-in-parameters models. This work applies VB method to the nonlinear fuzzy filters without using Taylor expansion for a linear approximation of some nonlinear function. It is assumed that the nonlinear parameters (i.e. antecedents) of the fuzzy filter are deterministic while linear parameters are stochastic. The VB algorithm, by maximizing a strict lower bound on the data evidence, makes the approximate posterior of linear parameters as close to the true posterior as possible. The nonlinear deterministic parameters are tuned in a way to further increase the lower bound on data evidence. The VB paradigm can be used to design an algorithm that automatically selects the most suitable fuzzy filter out of the considered finite set of fuzzy filters. This is done by fitting the observed data as a stochastic combination of the different Takagi-Sugeno fuzzy filters such that the individual filters compete with one another to model the data.
A Mixture of Fuzzy Filters Applied to the Analysis of Heartbeat Intervals
This study provides a stochastic modeling of the heartbeat intervals using a mixture of Takagi-Sugeno type fuzzy filters. The model parameters are inferred under variational Bayes (VB) framework. The model of the heartbeat intervals is in the form of a history-dependent probability density. The parameters, characterizing the heartbeat intervals probability density, include the estimated parameters of different fuzzy filters and may serve as the features of the heartbeat interval series. The features of the heartbeat intervals provide a description of the physiological state of an individual. A novelty of our analysis method is that the physiological state is predicted as a part of the features extraction procedure. This is done via deriving, using VB paradigm, an analytical expression for the posterior distribution that the observed heartbeat intervals have been generated by the stochastic model of the physiological state.