Rubiyah Yusof

Tuning of a neuro-fuzzy controller by genetic algorithm

Due to their powerful optimization property, genetic algorithms (GAs) are currently being investigated for the development of adaptive or self-tuning fuzzy logic control systems. This paper presents a neuro-fuzzy logic controller (NFLC) where all of its parameters can be tuned simultaneously by GA. The structure of the controller is based on the radial basis function neural network (RBF) with Gaussian membership functions. The NFLC tuned by GA can somewhat eliminate laborious design steps such as manual tuning of the membership functions and selection of the fuzzy rules. The GA implementation incorporates dynamic crossover and mutation probabilistic rates for faster convergence. A flexible position coding strategy of the NFLC parameters is also implemented to obtain near optimal solutions. The performance of the proposed controller is compared with a conventional fuzzy controller and a PID controller tuned by GA. Simulation results show that the proposed controller offers encouraging advantages and has better performance.

Hybrid Agent Tutor Model for e-Learning System with Robotics

In this paper, we propose Hybrid Agent Tutor Model (HATM) simulated with Robotics to enhance an e-Learning system for education field. The hardware robot agents in real world and software agents in virtual one were defined as the Hybrid Agent Tutors (HATs), and will perform interactive relationship so as to support learners to get more efficient educational contents. Based on this concept, we configured a simple prototype system to show flame work according on noble “Qualia” and “Awareness”, and tried to verify the efficiency of this system in e-Learning.

Feedback Preferences in Case-Base Construction for Intelligent Lab Tutor

Development of e-learning tools has been researched widely since 1900’s. Virtual laboratory (VL) has been one of the tools implemented, and it has given a great impact on student’s learning. However, the learning process become more challenging, as the students perform the experiments without any guidance from a human tutor. Consequently, this paper proposes the development of a case-based Intelligent Lab Tutor (ILT) to assist the learning process. The ILT is expected to provide scaffolding to students during their experiment sessions using the virtual lab. We intend to adopt student preferences as the initial case-base for the ILT. A questionnaire is distributed to a group of engineering students, and analyzed to identify students’ preferences in feedback type and frequency, given their performance in the experiment. The case representation is conceptualized at the end of this study.

Fault diagnosis of power transformer using optimally selected DGA features and SVM

Dissolved Gas Analysis (DGA) is an established method for detecting and predicting faults contained in power transformers. Support Vector Machine (SVM) has been actively applied to classify faults using historic DGA data. However, redundant and irrelevant features can reduce SVM classification performance. Therefore, this study proposes the use of GA-SVM wrapper to eliminate these features and select optimal features from DGA dataset to be used as inputs to SVM. GA-SVM wraps Genetic Algorithm (GA) around SVM, meaning that the estimated accuracy of SVM becomes the fitness function for each of the subsets found or generated by GA. Using these optimal features, SVM is trained and tested using two different datasets. The accuracies of SVM learned on the full set of features and that learned on the selected subsets by GA are compared using two real-world DGA datasets. Experimental results show that SVM performs better using optimal DGA subset than the whole dataset. It can be concluded that the proposed method which combines GA-SVM and SVM eliminates redundant features and improves SVM performance in classifying transformer fault based on DGA data.

Simple model following control design methods for a stable and an unstable plant

In this paper, we propose simple model following control design methods for a stable and an unstable plant. Each method is based on the pole placement using model parameters. In simulation studies, we show various cases including a nominal plant and the plant with a modeling error for a stable plant, an unstable plant, and a DC motor. These results show that the proposed methods have been superior performances.

Analysis of prediction methods for swarm robotic — In the case of oil spill tracking

Discharge of liquid petroleum on water surface, known as oil spill, frequently happen as result of an offshore well vessels failures or transportation accidents. In order to efficient treatment of these environmental catastrophes and diminishing the spots, the information about precise location and potential situation of the oil spill in future is significantly useful. Therefore, in this research the focus is to track and predict the spread of the slicking oil. More precisely, an algorithm of swarm robotic is proposed to engage the synchronized fuzzy controlled robots, in order to pursue the boundaries of an oil spill which is influenced by environmental forces such as wind and wave currents. Three methods of prediction such as linear extrapolation, Spline extrapolation and linear regression are applied and analyzed in this purpose.