National Taiwan University of Science and Technology

About: National Taiwan University of Science and Technology is a education organization based out in Taipei, Taipei, Taiwan. It is known for research contribution in the topics: Fuzzy logic & Control theory. The organization has 16288 authors who have published 21577 publications receiving 426294 citations. The organization is also known as: Taiwan Tech & Taiwantech.

FPGA-Based Intelligent-Complementary Sliding-Mode Control for PMLSM Servo-Drive System

Abstract: A field-programmable gate array (FPGA)-based intelligent-complementary sliding-mode control (ICSMC) is proposed in this paper to control the mover of a permanent magnet linear synchronous motor (PMLSM) servo-drive system to track periodic-reference trajectories. First, the dynamics of the field-oriented control PMLSM servo drive with a lumped uncertainty, which contains parameter variations, external disturbances, and nonlinear-friction force, is derived. Then, to achieve the required high-control performance, the ICSMC is developed. In this approach, a radial-basis function-network (RBFN) estimator with accurate approximation capability is employed to estimate the lumped uncertainty directly. Moreover, the adaptive-learning algorithms for the online training of the RBFN are derived using the Lyapunov theorem to guarantee the closed-loop stability. Furthermore, the FPGA chip is adopted to implement the developed control and online learning algorithms for possible low-cost and high-performance industrial applications using PMLSM. Finally, some experimental results are illustrated to show the validity of the proposed control approach.

Toward a hybrid data mining model for customer retention

Abstract: The prevention of subscriber churn through customer retention is a core issue of Customer Relationship Management (CRM). By minimizing customer churn a company maximizes its profit. This paper proposes a hybridized architecture to deal with customer retention problems. It does so not only through predicting churn probability but also by proposing retention policies. The architecture works in two modes: learning and usage. In the learning mode, the churn model learner seeks potential associations from the subscriber database. This historical information is used to form a churn model. This mode also calls for a policy model constructor to use the attributes identified in the churn model to divide all 'churners' into distinct groups. The policy model constructor is also responsible for developing a policy model for each churner group. In the usage mode, a churn predictor uses the churn model to predict the churn probability of a given subscriber. When the churn model finds that the subscriber has a high churn probability the policy model is used to suggest specific retention policies. This study's experiments show that the churn model has an evaluation accuracy of approximately eighty-five percent. This suggests that policy model construction represents an interesting and important technique in investigating the characteristics of churner groups. Furthermore, this study indicates that understanding the relationships between churns is essential in creating effective retention policy models for dealing with 'churners'.

Hexagonal microlens array modeling and fabrication using a thermal reflow process

Abstract: A mathematical model for designing and fabricating a hexagonal microlens array using a thermal reflow process was developed in this study. The experimental results proved that a hexagonal microlens array could be produced without a gap at each microlens periphery. A hexagonal microlens array with a higher fill factor was successfully produced. In this experiment, hexagonal photoresist columns were formed onto a silicon substrate made using a lithographic process. The hexagonal pattern was laid out in an ortho-triangle on a PET (polyethylene terephthalate)-based mask. Using precise temperature and time control during the thermal reflow process, a hexagonal microlens array with lateral honeycomb geometry was formed from the melted photoresist flowing outward simultaneously and uniformly. The surface tension effect transformed the photoresist column surface into a spherical profile. The error in the fabricated microlens characteristics was within ±3% between two theoretical models used to predict the photoresist column thickness and actual thickness. This model is feasible for fabricating various sized hexagonal microlens arrays.

A context-aware ubiquitous learning approach to conducting scientific inquiry activities in a science park

Abstract: Fostering students' scientific inquiry competence has been recognised as being an important and challenging objective of science education. To strengthen the understanding of science theories or notations, researchers have suggested conducting some learning activities in the field via operating relevant devices. In a traditional in-field scientific inquiry activity, the teacher usually lets the students operate the devices on their own after demonstrating the operational procedure. With such an approach, the students are likely to suspend the practice when they encounter problems; moreover, it is difficult for the students to connect what they have learned from the textbooks with the field practice. To deal with this problem, this study presents a context-aware ubiquitous learning system with sensing technology to detect and examine the real-world learning behaviours of students, such that personalised learning guidance and feedback can be provided; moreover, the students' experiences of operating those scientific devices, such as solar power equipment or the constellation simulators, can be conjunct to the knowledge learned from the textbooks. The experimental results from a science course of an elementary school show that this innovative approach is able to improve the learning achievements of students as well as enhance their learning motivation.

DEA and ranking with the network-based approach: a case of R&D performance

Abstract: This study enhances the network-based approach, which is a novel method to increase discrimination in data envelopment analysis. The enhancements include removing the bias caused by a scale difference among organizations and highlighting the approach's ability to identify the strengths and weaknesses of each organization. The former makes the approach applicable to both the constant returns of scale (CRS) and the variable returns of scale (VRS) models. The network-based approach applies the centrality concept developed in social network analysis to discriminate efficient decision making organizations as determined by standard data envelopment analysis (DEA). More specifically, the results of data envelopment analysis are transformed into a directed and weighted network in which each node represents a decision making organization and the link between a pair of node represents the referencing relationship between the pair. The centrality value for each efficient organization provides the base for discrimination and ranking. This network-based approach suggests aggregating DEA results of different input/output combinations such that the merits of each organization under various situations can be considered. The final ranking of this approach favors organizations that have their strengths evenly spread and tends to screen out specialized efficient organizations. As a real world example, the approach is applied to evaluate and rank the R&D (research and development) performance of Taiwan's government-supported research institutes. The cross-organizations and within-organization strengths for each efficient research institute are identified after applying the approach. A two-stage R&D evaluation model separates the R&D process into the technology development and technology diffusion stage. The resulting performance map differentiates the research institutes into four categories—Achievers, Marketers, Innovators, and Underdogs.