M. I. Khairir

Vision-Based Mobile Robot Navigation Using Image Processing and Cell Decomposition

In this paper, we present a method to navigate a mobile robot using a webcam. This method determines the shortest path for the robot to transverse to its target location, while avoiding obstacles along the way. The environment is first captured as an image using a webcam. Image processing methods are then performed to identify the existence of obstacles within the environment. Using the Cell Decomposition method, locations with obstacles are identified and the corresponding cells are eliminated. From the remaining cells, the shortest path to the goal is identified. The program is written in MATLAB with the Image Processing toolbox. The proposed method does not make use of any other type of sensor other than the webcam.

A Weighted Genetic Algorithm Based Method for Clustering of Heteroscaled Datasets

This paper introduces a weighted genetic algorithm (GA) based clustering method for datasets with differently scaled dimensions. Several types of synthetic two dimensional scatter data were clustered using the typical k-means clustering method. The weighted GA-based clustering method was developed to address the problem of clustering data with differently scaled (heteroscaled) dimensions. Cluster analysis results obtained from using this method was compared to the results produced from the application of the traditional k-means clustering. By introducing weights in the fitness evaluation component of the meta-heuristic search method, a more efficient clustering of heteroscaled data was produced. In real applications, this method can be used in cluster analyses of scatter data with significantly different scales in dimensions, such as kurtosis versus fatigue damage relationship scatter data.

Optimising Real-Time Performance of Genetic Algorithm Clustering Method

This paper presents the optimisation of real-time performance of the genetic algorithm clustering method. This performance optimisation concerns the population diversity and limitation and is based on actual runtime of the algorithm. A real-time ticker is incorporated into the algorithm for actual runtime measurement. For population diversity and limitation, a controlled k-means analysis is performed on the population of solutions to determine its diversity. Achieving a less diverse population in less amount of time without sacrificing the accuracy of the algorithm will help reduce the time-complexity of the algorithm, thus opening up the potential for the algorithm to cluster data in higher dimensions. Results from this study will be used for improving the method of clustering fatigue damage features of automotive components using genetic algorithm based methods.

The Development of Validation Technique in Variable Amplitude Loadings Strain Repetitive Data Collection

In any experiment, there is a need to verify the reliability of the collected data. One of the solutions is to measure the data repetitively. It will lead to measurement of inconsistency that exists in the data. Similar things also happen in variable amplitude (VA) loading strain data collection where the data also need to be measured repetitively in order the collected data is reliable. The main objective of this study is to validate the reliability of collected VA loading strain data. Two techniques will be used for identifying the similarity in the pattern for strain signal which was measured repetitively. The probability distribution function (PDF) and power spectral density (PSD) diagrams of the strain data were used as the main tool to construct profile plots for the case study data. Then, each profile plot will be compared to each other in order to identify any similarities that exist in the case study data. For the purpose of the study, a set of nonstationary VA loadings strain data that exhibits random behaviour was used. This random data was measured in the unit of microstrain on the lower suspension arm of a car. The data was repetitively measured for 60 seconds at the sampling rate of 500Hz, giving 30,000 discrete data points. The distribution of the collected data was analysed using the PDF and PSD. Based on the analysis, it was found that PSD can produce better results at identifying the similar features that exist in the profile plot compared to PDF.