Min-Soeng Kim

Evolving Compact and Interpretable Takagi–Sugeno Fuzzy Models With a New Encoding Scheme

Developing Takagi-Sugeno fuzzy models by evolutionary algorithms mainly requires three factors: an encoding scheme, an evaluation method, and appropriate evolutionary operations. At the same time, these three factors should be designed so that they can consider three important aspects of fuzzy modeling: modeling accuracy, compactness, and interpretability. This paper proposes a new evolutionary algorithm that fulfills such requirements and solves fuzzy modeling problems. Two major ideas proposed in this paper lie in a new encoding scheme and a new fitness function, respectively. The proposed encoding scheme consists of three chromosomes, one of which uses unique chained possibilistic representation of rule structure. The proposed encoding scheme can achieve simultaneous optimization of parameters of antecedent membership functions and rule structures with the new fitness function developed in this paper. The proposed fitness function consists of five functions that consider three evaluation criteria in fuzzy modeling problems. The proposed fitness function guides evolutionary search direction so that the proposed algorithm can find more accurate compact fuzzy models with interpretable antecedent membership functions. Several evolutionary operators that are appropriate for the proposed encoding scheme are carefully designed. Simulation results on three modeling problems show that the proposed encoding scheme and the proposed fitness functions are effective in finding accurate, compact, and interpretable Takagi-Sugeno fuzzy models. From the simulation results, it is shown that the proposed algorithm can successfully find fuzzy models that approximate the given unknown function accurately with a compact number of fuzzy rules and membership functions. At the same time, the fuzzy models use interpretable antecedent membership functions, which are helpful in understanding the underlying behavior of the obtained fuzzy models

Designing a robust adaptive dynamic controller for nonholonomic mobile robots under modeling uncertainty and disturbances

Abstract The main stream of researches on the mobile robot is planning motions of the mobile robot under nonholonomic constraints. Much has been written about the problem of motion planning under nonholonomic constraints using only a kinematic model of a mobile robot. Those methods, however, assume that there is some kind of a dynamic controller that can produce perfectly the same velocity that is necessary for the kinematic controller. Also there is little literature on the robustness of the controller when there are uncertainties or external disturbances in the dynamical model of a mobile robot. In this paper, we proposed a robust adaptive controller that can achieve perfect velocity tracking while considering not only a kinematic model but also a dynamic model of the mobile robot. The proposed controller can overcome uncertainties and external disturbances by robust adaptive technique. The stability of the dynamic system will be shown through the Lyapunov method.

Adaptive neurofuzzy controller to regulate UTSG water level in nuclear power plants

A data-driven adaptive neurofuzzy controller is presented for the water-level control of U-tube steam generators in nuclear power plants. This neurofuzzy controller is capable of learning the control action principles from the data obtained using other methods of automatic or manual control. There are four inputs in the neurofuzzy system, yet only eighty fuzzy rules involved. Therefore, the fuzzy system is versatile and moderately compact. The versatility is due to the higher input space dimension that helps to learn more control principles. The compactness is due to the number of rules being not too many. A 10-h evaluation trial of the trained fuzzy controller demonstrated its capability in regulating the water level under random disturbances and reference level changes.

Control of mobile robots using RBF network

This paper deals with handling the unknown factors, such as the external disturbance and the unknown dynamics, for a mobile robot control. We propose a RBF network based controller for compensate for them. The stability of the proposed controller is proven by using Lyapunov function. To show the effectiveness of the proposed controller, several simulation results are presented. Through the simulations, we show that the proposed controller can overcome the modelling uncertainty and the disturbances. Also the proposed RBF controller outperforms the previous works from the viewpoint of computation time, which is a crucial fact for a real-applications.

Heterogeneous local model networks for time series prediction

The approaches of local modeling have emerged as one of the promising methods of time series prediction. By use of the divide-and-conquer method, local models can exploit state-dependent features to approximate a subset of training data accurately. However, the generalization performance of local model networks is subject to the proper selection of model parameters. In this paper, we present a new method for local model construction for the noisy time series prediction. The proposed method uses the principal component analysis (PCA) and cross-validation technique to construct an optimal input vector for each local model. A heuristic learning rule is also proposed to update the mixture of experts network structure, which determines the confidence level of local prediction model. The proposed method has been tested with noisy Mackey-Glass time series and Sunspot series.

Multi-objective walking trajectories generation for a biped robot

The generation of the optimal walking pattern is an important question for a biped robot to keep walking stably. This paper is proposed for generating the walking patterns resulted in best performance of the biped robot using multiobjective evolutionary algorithm. We formulate a trajectory generation problem as a multi-objective optimal problem. We obtain all Pareto-optimal solutions on the feasible solution region for various walking pattern generation of a biped robot in EA simulation.

Control of a nonholonomic mobile robot using an RBF network

This article deals with handling unknown factors, such as external disturbance and unknown dynamics, for mobile robot control. We propose a radial-basis function (RBF) network-based controller to compensate for these. The stability of the proposed controller is proven using the Lyapunov function. To show the effectiveness of the proposed controller, several simulation results are presented. Through the simulations, we show that the proposed controller can overcome the modelling uncertainty and the disturbances. The proposed RBF controller also outperforms previous work from the viewpoint of computation time, which is a crucial fact for real-time applications.

Evolutionary optimization of fuzzy models with asymmetric RBF membership functions using simplified fitness sharing

A new evolutionary optimization scheme for designing a Takagi-Sugeno fuzzy model is proposed in this paper. To achieve better modeling performance, asymmetric RBF membership functions are used. Penalty function is proposed and used in the fitness function to prevent overlapping membership functions in the resulting fuzzy model. The simplified fitness sharing scheme is used to enhance the searching capability of the proposed evolutionary optimization algorithm. Some simulations are performed to show the effectiveness of the proposed algorithm.

Incremental Hyperplane-based Fuzzy Clustering for System Modeling

In this paper, a new incremental hyperplane-based fuzzy clustering method to design a Takagi-Sugeno-Kang (TSK) fuzzy model is proposed. Starting from no rule, it generates clusters based on input similarity and distance from the consequent hyperplane incrementally. Membership functions (MFs) are defined with statistical means and deviations of partitioned data. With this configuration, the obtained clusters reflect the real distribution of the training data properly. The training equations are changed to recursive forms in order to be applied in incremental framework. Some heuristic techniques to guarantee the initial training of each local submodel is used. In order to reduce the dependency on the order of training data, merge step is performed. Merge step is not only important for keeping rule bases compact and interpretable, but also provides the robustness to noise. Some simulations are done to show the advantages and performance of the proposed method.

Robust model reference adaptive control of underactuated robot manipulators

A robust control scheme, overcoming the uncertainty in an underactuated robot manipulator, is proposed based on the sliding mode and MRAC (model reference adaptive control) schemes. By introducing the model reference adaptive technique and robust control algorithm, the dynamic response of each joint of underactuated manipulators can be pre-determined without exact knowledge of the system parameters. To show the effectiveness of the proposed algorithm, simulations for a 2-link underactuated robot with 1 fault joint are done.