2011 THE UNIVERSITY OF AIZU ALL RIGHTS RESERVED
Principal Investigator: Kazuyoshi Mori
Title: Characterization of Multiple-Stage Stabilization for Linear Systems
Amount: ¥724,000
In this project we have investigated: (i) characterization of alternative two-stage stabilization, (ii) characterization of alternative two-stage stabilization in total feedback system.
For (i) and (ii), we have obtained the parametrization of stabilizing controllers based on the feedback system. They are given as forms modified the Youlaparametrization form. We have also obtained other alternative two-stage stabilizations. By the result of (ii), we can obtain all stabilizing controller of this two-stage stabilization.
Principal Investigator: Yong Liu
Title: Optimization of Photonic Band Structure by Fast Evolutionary Programming
Amount: ¥658,000
Photonic crystals (PhCs) are composed of periodic high and low dielectric constant nanostructures. PhCs affect the propagation of electromagnetic waves in a similar way that periodicity of a semiconductor crystal affects the motion of electrons through specifying allowed and forbidden electric energy bands. The periodically electromagnetic media also possesses photonic band gaps, in which light within certain ranges of frequency cannot propagate through. Such photonic band gaps make photonic crystal be attractive optical materials for controlling and manipulating the flow of light. One-dimensional photonic crystals are already in widespread use in the form of thin-film optics with applications ranging from low and high reflection coatings on lenses and mirrors to color changing paints and inks. Three major problems have been addressed in the research. They are 1) randomness analysis in the design of thin-films; 2) optimization of the structures of thin films in the design of dichroic laser beam combiners; 3) analysis of structures of one-dimensional PhCs in the implementations of optical switching devices.
The results of this project were presented in one international journal, and three referred international conference papers. Two volumes of edited proceedings related to this project were published by Springer, including Lecture Notes in Computer Science (LNCS 6382) and Communications in Computer and Information Science (CCIS 107).
Principal Investigator: Victor I. Ryzhii
Title: Grapheme-New Material for XXI Century Electronics: Theoretical and Numerical Studies of Novel Grapheme Deices (III)
Amount: ¥1,091,000
A novel terahertz/infrared detector of electromagnetic radiation based on NonBernal stacked graphene layers has been proposed and evaluated using the developed models. The proposed and studied device surpasses many already used photodetectors, particularly at elevated temperatures (T = 300K).
The results, of the project have been published, in particular, in
M.Ryzhii, V.Ryzhii, T.Otsuji, V.Mitin, and M.S.Shur, "Electrically induced p-i-n junctions in multiple graphene layer structures", Phys. Rev. B, vol. 82, 075419 (2010),
V.Ryzhii, M.Ryzhii, N.Ryabova, V.Mitin, and T.Otsuji, "Terahertz and infrared detectors based on graphene structures", Infrared Physics and Technology, vol. 54, 302 (2011),
and presented at conferences and seminars.
Principal Investigator: Irina I. Khmyrova
Title: Analytical study and IsSpice simulation of distributed systems for teraherts, resonant sensor and bio-medical applications
Amount: ¥1,112,000
The goal of the Project was to study the impact of fringing effectcs on the performance of terahertz deevices with two-dimensional electron gas channel periodically modulated by interdigitated gate. Analytical model for evaluation of spatial distribution of sheet electron density has been developed. Resonant frequences of plasma oscillations in the 2DEG channel of such structures have been evaluated in the ideal and realistic situations. It is shown that the increase in the number of the gates results in the reduction of the separation between adjacent harmonics which can be one of the factors responsible for spectrum broadening observed experimentally. The proposed model can be useful for interpretation of experimentally obtained data, in particular, in the evaluation of the harmonic number.
Principal Investigator: Maxim V. Ryzhii
Title: Computer simulations of biophysical phenomena in living tissue
Amount: ¥375,500
Many problems in cardiac electrophysiology are not studied in detail yet. Investigations on human ventricular myocardium are still very limited due to difficulties in accessing in vivo data. Alternative methods of study such as computer modeling of electrical and magnetic features of the heart and brain are of great importance. Utilization of modern simulation approaches allows obtaining clear insight in many abnormal situations in cardiac electrophysiology as well as additional novel methods of their treatment. Simulations of heart based on the solution of reactiondiffusion equations together with precise models with realistic shapes, small mesh sizes and time steps require utilization of high-performance computing systems.
The purpose of the research was computer modeling study of computationally demanding problems arising in cardiac electrophysiology and soft tissue homeostasis and carcinogenesis. The main goal of the project was obtaining cardiac electrophysiological characteristics under abnormal conditions as well as abnormal soft tissue growth characteristics. We considered of biophysical models of the systems under study, modified freely available software package, and performed numerical simulations utilizing high-performance computer workstation. As a result of the project, two papers at international conferences were presented.
Principal Investigator: Yukihide Kohira
Title: Design Automation Methodology for LSI chips with Programmable Delay Elements
Amount: ¥820,000
In resent LSI chips, process variations increase significantly because of the progress of the process technology. The process variations significantly cause delay variations such as gate delay variations or wire delay variations, and delay variations affect the performance and the yield of VLSI chips. In this project, deskew technique was applied to establish the design method for LSI chips with high tolerance to delay variations. Deskew is the adjustment of the PDEs (programmable delay elements) after the fabrication of LSI chips. In the proposed design flow, the chip with PDEs was implemented by a commercial CAD tool. Experiments showed that we can obtain the circuit with high tolerant to delay variations without increasing the power consumption.
Principal Investigator: Jungpil Shin
Title: A Study on Pen Computing
Amount: ¥1,027,000
The purpose of our project, "A Study on Pen Computing," is the development of Pen Computing algorithm based on the Handwriting with high performance. We have developed the algorithms ad systems for the following goal.
To develop a handwriting analysis system with Wii enabling verifying a signer correctly (error rate less than 5%) and recognizing characters correctly (error rate less than 4%).
To achieve the robust recognition for Waseda style Japanese shorthand.
To develop a character synthesis system, enabling synthesizing a character when stroke connections are not stable and the probabilistic generation of stroke connection.
To develop a signature verification system, enables decrease error rates using the synthesizing model and the synthesized signatures, verifying a signer correctly (error rate less than 5%) even when the signer is intoxicated, and evaluate user-input signature automatically and reject to register to the system if the complexity is too low.
To develop a watercolor drawing simulation system that realizes the realistic of coloring style.
Principal Investigator: Daming Wei
Title: A new application of brain-computer interface to rehabilitation for stroke patients
Amount: ¥1,300,000
Recently, increase in stroke patients associated with changes in lifestyle and aging is a big social problem. This desease can lead to various disabilities, particularly know for its many movement disorders. However, if done early after onset of stroke, some patients have been shown in recent studies to a large extent able to recover motor function by rehabilitation, but also to promotions of brain plasticity.
The motor learning is the basic mechanism in traditional rehavilitation. Learning and exercixe, as the directive came from the motor cortex and spinal cord, contract the muscle repetedly to form a loop back to the sensory area of the feedback information and to improve the motor function. However, dor to brain damage in stroke, the neural loop is interrupted and cannot achieve the desired motion. In this project, we expect to establish a new loop between the brain and muscle by application the brain-computer interface (BC) technology. An artificial loop starting from orders from the motor cortex of the brain to muscle nerves in helps with FES (Functional Electrical Stimulation) and a robotic arm is intended to facilitate the recovery of motor function.
Principal Investigator: Wenxi Chen
Title: Deep Body Temperature Measurement Using Multi-Heart-Flux Method
Amount: ¥1,041,000
"Deep Body Temperature Measurement Using Multi-Heart-Flux Method" This study is to clarify two issues: the first is to validate the method theoretically through modeling the heat transfer process of a multi-heat-flux probe in measuring deep body temperature; the second is to acquire a detailed quantitative description of the dynamic mechanism of the probe in order to improve measurement performance in terms of accuracy and response time.
Principal Investigator: Incheon Paik
Title: Design of Customizable Engine Framework for Automatic Web Service Composition
Amount: ¥1,003,000
In this project, an entire composition engine framework with functional-scalable workflow generation, automatic properties transformation, and seamless operation has been developed. The five issues have been addressed for final goal of the project: functional-scalable workflow management, automatic discovery of functional and non-functional properties, automatic development of workflow from request, entire architecture for composition engine, prototype for proof of concept. The fourstaged composition architecture was used for composer with general domains. A software development tool, we call it NASCDK (Nested Automatic Service Composition Development Kit) has been announced. A demonstration using trip example has been introduced.
Principal Investigator: John Brine
Title: Lexical Support for extensive reading on eReaders
Amount: ¥1,119,000
This research investigated the development of an electronic eReader to support second language acquisition through extensive reading. Extensive reading involves strategies to increase student reading at levels appropriate to student ability. According to the literature on second language reading, ESL students acquire more vocabulary when they have access to computer-mediated glosses and dictionaries. We developed an eReader prototype using Wikification technologies to extend exploration of lexical and conceptual knowledge resident in text databases. The prototype was implemented for Apple iPad using application development tools such as Xcode. We also examined integration of the eReader into a cloud-based system. This research was important in establishing a basis for continued research into the use of mobile devices for language learning applications.
Principal Investigator: Paul A. Lyddon
Title: Implementing an Online Video Discussion Tool for Asynchronous Collaboration and Exchange
Amount: ¥263,000
This project entailed the local installation and adaptation of the Online Language Environments (OLE) system, an open-source video discussion board originally developed at the University of Arizona to promote foreign language listening and speaking practice outside of class. After a number of interface, functionality, and backend modifications, the system was piloted in one section of English Speaking and Writing 2, for which the students completed two online speaking assignments. On an end-of-course survey, 75% of respondents evaluated the first assignment as either `somewhat' or `very' important for improving their speaking ability (mode = `somewhat'), and nearly 95% evaluated the second assignment similarly (mode =!Fvery !G). Moreover, 80% indicated that they thought OLE should continue to
be used in future speaking and writing classes, the most common suggestion for improvement being to use it for a greater number of assignments.
Principal Investigator: Shuxue Ding
Title: Sparse Representation of Non-Negative Signal
Amount: ¥889,000
In this research, we proposed and investigated a novel technology for blind source separation of non-negative sources based on the sparse representation. We invoked the non-negative matrix factorization (NMF) for solving the problem. NMF could be formulated in an optimization problem, with the non-negative constraints on the two matrix factors. !!In this technology, the uniqueness of the sparse representation can be dramatically enforced, which was a problem in the conventional methods. On the other hand, we can relax the condition on the independence of the sources in this problem setting. That is, we can even separate statistically dependent sources. Although the non-negativity is necessary, the technology still has very broad potential applications, such as, image processing, spectral unmixing, biomedical processing, etc.
Principal Investigator: Taro Suzuki
Title: Implementation of communication system of lunar and planetary exploration data besed on XPEF(XML-based Planetary Exploration data Format)
Amount: ¥906,000
This project is a follow-up to our project in the last fiscal year subsidized by Competitive Research Funding. In FY2009 we proposed a novel data format for lunar and planetary data format, called XPEF, and implemented a transformation tool from PDS, a widlely used but old-fashoined data format for lunar and planetary data format, to XPEF for a specific exploration mission.
In FY2010, we needed to adapt our transformation tool to the messaging network, a network that enables transmission of data in transparent way by providing transfomation of data suitable for their recipients. In order to achive demand-driven transformation of different PDS data for various exploration missions under the messaging network, we designed a generator of transformation tools from PDS to XPEF to support multiple PDS data obtained by various exploration missions. It receives template Java source files, which remain unchanged through all the missions, and a metadata information for a specific mission. The metadata information is a XML document, where its elements are the ones specified in the XML schema for XPEF format, called XPEF elements, and the value (a string) of each element is the name of PDS data item corresponding to the XPEF element. Thus, the metadata information gives the correspondence between PDS data items appearing in PDS data obtained by a specific mission and XPEF elements. Every XPEF element in the metadata file has an attribute, specific to the description of metadata information, that specifies how the value of the corresponding PDS data item is described in a PDS file.
Principal Investigator: Tsuneo Tsukahara
Title: A High-Precision Complex Quadrature Modulator for A Multi-bant Wireless Transceiver
Amount: ¥869,000
Recently the demand for wireless systems has been rapidly growing. However, radio-wave resources are limited and invaluable especially in these days. Therefore, software-defied radios (SDRs) and cognitive radios (CRs) can be the key to greatly improving frequency-spectrum efficiency. SDRs demand flexibility and reconfigurability in RF (Radio Frequency) circuits. Therefore, a spectrum-efficient wireless transceiver architecture is indispensable. In this RD, we proposed a multiband wireless transceiver using a high-precision complex quadrature modulator and a spectrum-analyzing receiver. In recent years, multi-level modulations such as Quadrature Amplitude Modulation (QAM) are used in Wireless LANs and digital TVs, and will be used in the 4th-generation cell-phones. The QAM demands very small modulation errors of quadrature modulator LSIs with less than 1%. In 2010, we have proposed and designed the high-precision complex quadrature modulator suitable for direct-conversion transmitters of multi-band SDRs. Featuring the inherent correction mechanism of phase and amplitude errors, it can provide small modulation errors with a wide bandwidth. The impact of phase and amplitude errors was simulated using the single-sideband (SSB) modulation. Simulation parameters are based on the Rohm 0.18-um CMOS technology provided through the VDEC. The proposed quadrature modulator can provide the sideband rejection ratio (SRR) of more than 50dB, thereby being applicable to QAM systems. Because we need a high supply-voltage of more than 1.8 V, the next step is implementing low-voltage and low-power operations.
Principal Investigator: Robert H. Fujii
Title: Biologically Insired Optimization Applied to Neural Motor Learning, Algorithm Engineering, and Artificial Wind Tunnel Experiments Amount: ¥1,254,000
AI optimization of vehicle shapes to reduce aerodynamic resistance.
This is a generally applicable energy-saving method for reducing drag and turbulence caused by the flow of a fluid and/or a gas over surfaces.
AI optimization (improvement) of the strategic game of chess.
This has general applications for any game of strategy that requires a decision tree style thinking and a high level of situational awareness.
A soft light-weight actuator.
The fabrication of a soft actuator that can operate in air at a low voltages . Such an actuator can convert electrical energy directly into mechanical energy without the use of intervening coils or gears and can be used for microscopic (e.g. micro tweezers) and macroscopic (e.g. robot arm actuator) applications.
Results
AI optimization:
Reducing wind resistance of the shape of a truck by a significant margin compared to current technologies seen on the roads.
After up to 5,000 hours of distributed computation time for a single genetic algorithm optimization, an approximately month-long genetic algorithm optimization procedure with over 100 generations with a population of 16-20 candidate solutions followed. We were then able to obtain an optimized simple car-model which (up to practically necessary modifications) can be seen on today's roads and, surprisingly, an optimized model for the air-shield on top of a driver's cabin of a truck with a shape which appears to be a new design. Both shapes found by the genetic algorithm described in this work beat elementary geometric shapes such as triangular shapes by a wide margin in a simulation of aerodynamic resistance. In particular, the new shape for the air-shield on top of a driver's cabin of a truck beats parabolic shapes of such windshields which can often be seen on today's roads by a wide margin, 20%. The application of the latter fact could yield energy savings in the transportation sector in the future. We demonstrate at the same time that through use of low-cost hardware and open-source free software tools, the optimization of aerodynamic or hydrodynamic performance of products by computer means is economically feasible in many situations such as external shapes of vehicles or machinery that transfers gases or fluids.
Improvement of GNU-chess software by 5%.
After up to 17,000 hours of distributed computation-time for a single optimization on a network of 17 LINUX workstations, the genetic algorithm finds a chess program that shows a moderate performance improvement compared with the original GNU-chess program. What appears to be new in the approach presented here are: (a) a brute-force optimization using a mentor rather than a co-evolutionary approach is actually carried out with contemporary PC hardware, and (b) optimizations for playing white and black are carried out separately which seemingly has not been attempted by other means before.
A preparatory GA-experiment already increased the playing power of an altered GNU-chess program by a small margin playing white (fitness 54.45% wins to 55.8%), but also showed the necessity to separately optimize the evaluation-coefficients for positions within the GNU-chess program for playing white and for playing black. In such an optimization for playing black, a moderate improvement for the evaluation-coefficients of the GNU-chess program playing black was found with a refined version of the distributed genetic algorithm (fitness 45.6% to 46.62%).
We note that actually fitness 59.2% over 40 games was obtained for the best creature in the genetic algorithm simulation optimizing the GNU-chess program playing . However, this value was not maintained in a separate simulation over several thousand games.
Soft Actuator:
A soft actuator was successfully fabricated using polymers. Fabricated actuator properties: size 1 cm X 2 cm, operates in air with+/− 2 V step input, Current +/−
15 mA, tested sample size 0.5 cm X 1.5 cm, test operating frequency 0.2 Hz., life cycle: not measured.
The bending of the soft actuator in response to the applied electric field occurs as a result of shape changes in the electrode layers. The anions (positive charge) and cations (negative charge) in the electrolyte layer are transferred to the cathode and anode layers respectively. This ion transportation mechanism results in the swelling at the cathode electrode and shrinkage at the anode electrode. As a consequence, the actuator bends in the direction of the anode.
Soft and lightweight actuators that can directly transform electrical energy to mechanical energy can have applications in areas such as robotics, tactile sensing, Braille character generation, prosthetic devices, medical devices, and micro-surgical devices.
Principal Investigator: Konstantin Markov
Title: Research and Development of basic technology for virtual smart meeting room
Amount: ¥381,000
The purpose of this project is to create a new type of meeting room environment based on the latest developments in Spoken Language Technology (SLT) and virtual reality research. Automatic indexing and annotation of multimedia (audio, video, text, etc.) data is an actively researched area. Data are usually collected from a life meeting in a meeting room which requires complicated and expensive equipment. This can be avoided when the meeting takes place in virtual reality. Each participant communicates through a computer which makes the acquisition of all the data simpler. This technology allows for segmentation of the meeting into speaker turns, transcription of speaker words as well as summarization of the meeting discussions and report generation.
Principal Investigator: Shigeru Kanemoto
Title: Automated Quantitative Skill Evaluation Technique for Nuclear Power Plant Operator Training
Amount: ¥979,000
The purpose of this research is to improve the trainee's skill evaluation system in the Boiling Water Reactor Operator Training Center (BTC) by utilizing computer vision technology to get necessary information from the video of training. The goal in AY2010 was to develop an image processing system to automatically extract information for trainee's skill quantitative evaluation.
In AY2010, we had enhanced our experimental image processing system designed in AY2009, by applying particle filter algorithm to automatically detecting trainee!G s position inside the training room. The experiment results were summarized and submitted to two international conferences and one Journal.
Principal Investigator: Kiyomi Okamoto
Title: Development of the University of Aizu Corpus: Corpus of Research Auticles Designed as Major Journals by the University
Amount: ¥375,000
According to the literature on English for specific purposes, research articles (RAs) are one of the typical genres, or particular types of communicative events with particular communicative purposes recognized by the users - or discourse community, in an academic discipline. One of the characteristics of a genre is its specific use of vocabulary. Although RAs in general share common features of lexis and discourse, there are differences among the fields of study. The University of Aizu designates a list of research journals as Major Journals. This list is divided by seven sublists. In this research project, a corpus of RAs from the Major Journals will be developed to analyze the linguistic features of the RAs in the Major Journal list. It is also our aim to investigate the differences in language use among the RAs in three sub-lists by dividing the corpus into seven sub-corpora, each representing an area of study in the list of Major Journals. The information obtained through the analyses will be indispensable for further development of language teaching at the University of Aizu as both readership and authorship of such journals contribute to the development of an academic discourse community, and inclusion and active participation at such communities should be one of their objectives in our students!G academic careers. The project produced a corpus which contains 45 research articles from nine prestigious journals in computer science and related fields. There are 434049 tokens and 16220 types in the corpus. A preliminary comparative study reveals that though the size of our corpus is much smaller, it has the similar range words contained compared to the computer science sub-corpus of PERC Corpus. In the future, this corpus will be an indispensable resource for developing teaching material and various examinations including entrance examinations.
Principal Investigator: Song Guo
Title: Modeling and Performance Evaluation of Cognitive Radio Networks with Multimedia Applications
Amount: ¥996,000
With the emergence of a variety of wireless multimedia applications, most of the usable radio spectrum has been densely allocated, which leads to the problem of spectrum scarcity worldwide. However, extensive measurement studies have indicated that the prime radio spectrum experiences significantly low utilization efficiency because of the current static spectrum allocation policy. Sparked by recent advances in cognitive radios, a new communication paradigm presents a possible solution to the spectrum inefficiency problem, which allows secondary users equipped with cognitive radios to opportunistically access unoccupied bands of licensed spectrum while limiting the interference perceived by the primary users. This networking paradigm is referred to as Cognitive Radio Networks (CRNs). In this project, we aim at developing QoS-aware MAC (Media Access Control) schemes for CRNs with heterogeneous multimedia applications. We also develop a unified analytical framework to evaluate the proposed QoS-aware MAC scheme in CRNs.
Principal Investigator: Tuan Anh Pham
Title: Design and Performance Evaluation of the next-generation ultra wideband optical-wireless home/home-access networks
Amount: ¥819,000
This research devotes to design and performance evaluation of the next-generation ultra-wideband home/home-access networks. The optical-wireless (OW) technology is proposed considering that it is a compelling solution (1) to overcome the RF spectrum scarcity and (2) to provide ultra-wideband wireless connectivity for home/home-access networks. The objective in this research focuses on novel modulation, multiplexing, coding and signal detection techniques to improve the performance of both home (indoor) and home-access (outdoor) OW links so that higher speed connectivity (up to Gb/s) for multiple users can be achieved.
Title: Digital Manga Drawing Stylization from Photographs
Amount: ¥725,000
Manga is a popular artistic media all around the world, especially in Japan, which is the origin of manga. Its distinguished fine drawing style requires intensive labor and time of manga artists to prepare. Although in recent years some artitst have already employed commercial software to speed up the production and reduce cost, such software usually produces simple halftoning or hatching results from photographs with are obviously inconsistent in style with traditional manga. This indicates the need to automate or digitally assists the manga production. The main objective of this project is to develop and depy novel technologies with intuitive Graphical User Interface (GUI) that facilitate manga auto-generation from real photographs.
Principal Investigator: Xin Zhu
Title: Study on Sleep Monitoring and Analysis System
Amount: ¥643,000
Sleep occupies a large part of an individual !Gs life and is essential for human beings to recover from exhaustion and maintain health. However, many cardiovascular diseases, such as hypertension, atherosclerosis, stroke, and coronary artery disease, may influence the quality of sleep (QOS) and lead to other severe symptoms and even sudden death during sleep. Ben-Dov et al. found that the heart rate (HR) measured during sleep and the absence of decrease in the HR during sleep were independently associated with all-cause mortality. Therefore, monitoring of the physiological signals during sleep is an important means of obtaining physiological information for the evaluation of health and diagnosis of cardiovascular and respiratory diseases. Polysomnography is a traditional method of monitoring sleep and is used for the diagnosis of sleep-related diseases in hospitals. However, the polysomnography test is expensive and requires expert intervention during the installation and monitoring procedures. Furthermore, subjects may experience difficulty in sleeping in a hospital and may not be as relaxed as they are at home. This decreases the reliability and accuracy of the polysomnography test. Furthermore, the data recorded during the polysomnography test are used for clinical diagnosis
and inconvenient for basic researches. Therefore, it is necessary to study and develop a platform and environment at the Biomedical Information Technology Lab for sleep monitoring and analysis through integrating the previously developed technology and methods.
In this research, we developed a platform and environment based on BIOTOP M36 for sleep monitoring, and studied related signal processing and analyzing methods for the analysis of sleep disorders such as sleep apnea. A paper related to this study was published in IEEE ISAC 2010 held in Tainan, Taiwan.
Principal Investigator: Yutaka Watanobe Title: Organizing and retrieving largescale data based on spatial allocation
Amount: ¥780,000
More flexible and efficient searching interfaces for large-scale data have been developed. Using the interfaces, users are able to find!Hnext items!Iin huge repositories which are similar, roughly similar in some properties as well as inspired items associated with a query item. Algorithms for space allocation of tags which associate the items have been analyzed. The algorithms have been evaluated through images from a huge image database which includes more than 1,000,000 items associated by 400,000 tags gathered from Flickr.
Principal Investigator: Takeaki Sampe
Title: Numerical experiments of regional weather phenomena
Amount: ¥953,000
Some kinds of weather phenomena, such as heavy rain/snow, occur on a scale of several kilometers, so that simulations of such phenomena are very difficult and require numerical modeling at a very high resolution. To do such numerical modeling of atmosphere, we used high-resolution regional atmospheric models with its boundary conditions constructed from output of a global atmospheric model. The research funding was used for a computer and a large disk storage system to handle huge data that are created in high-resolution simulations.
In FY2010, we performed numerical simulations with atmospheric models, called CReSS and WRF. An numerical experiment of heavy rain around Fukushima in August 2008 shows that the resolution of the numerical model affects not only the simulated rainfall amount but also the pattern and movement of the strong rain bands over Tohoku region. This finding highlights the imporatance of resolution in the simulation of a severe weather event.
Principal Investigator: Kohei Kitazato
Title: Development of Robotic Telescope System for Asteroid Search
Amount: ¥777,000
Robotic telescopes are telescopes that automatically observe the sky, process images and choose a subsequent strategy for observations. Such telescopes have the advantages of low operating costs with high operating efficiency and high scientific productivity. Their softwares developed so far, however, are specialized products,
and it is difficult to reuse them on different telescope systems. This project aim to develop an open-source software package to construct the robotic telescope system.
Principal Investigator: Debopriyo Roy
Title: Studying the Efficacy of Concept Maps in Technical Communication Courses: Understanding Text Applications and Interface Richness
Amount: ¥325,500
As part of this project, I argued for the importance of using specific kinds of concept maps, such as ladders, in a technical writing class offered for computer science majors. This project presented a strong argument that use of concept maps should be effectively integrated both during planning stages and as part of traditional document production techniques. Assignments and activities in a technical writing classroom should address the use of concept maps for conceptualizing, schematizing and presenting procedural information like software documentation, explaining laboratory reports etc. Using concept maps efficiently can lead to structural and functional conceptualization, visualization, representation and retention of complex information. A survey-based study reported in this project suggested that advanced students in an EFL technical thesis writing course in this computer science institute has enough expertise to understand the use and application of specialized concept maps for technical document production activities. Results indicate that students, on most part, understand specialized document-specific, task-specific applications and especially specific sentences in the domain of computer science and self-reports also indicate that they are reasonably confident about their ability to apply ladders in logical ways.
Principal Investigator: Hiroyuki Sagawa
Title: Structure and Reactions of Unstable Nuclei
Amount: ¥980,000
We studied structure and reactions of unstable nuclei by using microscopic models. Especially we focus charge exchange reactions to see the tensor correlation effect on Gamow-Teller states and Spin-dipole states. We found unique multipole dependence of tensor effect on Spin-dipole resonances which can be used to pin down triplet-even and triplet-odd parts of tenros interactions empirically. The results are published in several refereed journal papers.
Principal Investigator: Hitoshi Oi
Title: Resource management in virtualized systems
Amount: ¥909,000
System-level virtualization provides us with logically independent computing environments (virtual machines, or VMs) and system resources can be multiplexed so that each VM considers as if it has own instances of resources.
In this project, we investigate the resource demand and usage in typical and representative server applications and their sensitivity to the quality of service (QoS) metircs of the applications. We study the methodologies of resource management in the current virtualized systems and try to find better solutions.
Principal Investigator: Gennadiy Nikishkov
Title: Watar flow simulation on GPU
Amount: ¥811,000
Realistic water surface modeling with its evolution over time is important in computer graphics, especially in games. A common technique to produce realistic water surface animation is physically-based fluid simulation. In this research, we developed an algorithm for parallel integration of the damped wave equation on the graphics processing unit (GPU) using the NVIDIA CUDA programming model. The computational procedure includes integration of the damped wave equation and rendering the water surface using water height values defined at grid points. CUDA kernels have been developed for integration of the wave equation and for determination of the normal vectors at the grid points. Measurements of time for the water height integration and for the normal determination are performed by running CPU methods and GPU kernels. Considerable speedups of 17-25 are obtained for computational domains with ten thousand grid points and more. It is worth noting that domains with size of powers of two are characterizes by higher speedup values than domains of arbitrary sizes.
Principal Investigator: Ryuichi Oka
Title: Free-viewpoint 3D TV
Amount: ¥964,000
The objective of the project is a realization of real time 3-dimensional reconstruction of real world which is called Free-viewpoint TV (FTV). The system uses two fundamental techniques, namely, full pixel corresponding between two images and factorization for realizing 3-D image using a set of pixel corresponding data. The factorization is a well-known technique. The full pixel corresponding is obtained by our newly developed algorithm called two-dimensional Continuous Dynamic Programming (2DCDP).
The 2DCDP consists of a family of algorithms each of which is selected for a special application. The FTV uses an one algorithm called weak-spotting algorithm belonging to the family. The weal-spotting only requires a relatively small amount of memory and computation burden so that real time realization using a note PC becomes available.
Principal Investigator: Mohamed Hamada
Title: Enhanced Learning Style Index: Implementation and Application
Amount: ¥303,000
Learning preferences can help learners to find their most effective way to learn. It also can help teachers to adopt suitable learning materials for efficient learning. This project is concerned with the study, implementation, and application of a web-based learning style index. We also show a case study on the integration of learning style index into an adaptive and intelligent e-learning system.
Principal Investigator: Subhash Bhalla
Title: Design of New Query Language Interfaces for XML based Data Resources for Spatial data and Medical Informatics
Amount: ¥910,000
Web is an interactive medium for information exchange. In recent times many web-based applications exist in different domains which require the use of a query language interface. Personal health records (PHRs) are being commonly adopted in health information systems. Patient centeric portals, such as Google Health and Microsoft health vault are becoming popular. Ability to query over this data plays an important role for users. Similarly, agriculture knowledge repositories and biomedical information repositories have become available. The users need to query over large data. Most of the data resources use XML for storing data.
Many new applications also have geographic and spatial data components in XML form. There is a need to make spatial data query tools for semi-skilled users. On the support side, query languages exist for database management systems. These languages require knowledge of programming skills. The proposed research aims at developing easy to use higher-level query interfaces for large data repositories. This effort will increase the ability of users to query information for diverse data resources.
Principal Investigator: Emiko Kaneko
Title: Working memory training and oral proficiency in English as a second language
Amount: ¥734,000
The goal of this research was to find out whether students who undergo the training of reading aloud, repeating and shadowing English sentences would improve their speaking ability in terms of accuracy, fluency and complexity. It is believed that the reading-aloud and repeating skills are correlated with speaking because they tap into learners !Gworking memory, which conceivably helps them speak a second language (L2) better. Nonetheless, it is not certain how valid the correlation between working memory and speaking proficiency is for novice level learners. This research addressed this issue. Namely, it investigated whether training that was assumed to improve one !Gs working memory would influence speaking proficiency of novice level learners. This training method may improve one !Gs working memory, which in turn will have a positive effect on speaking an L2. The effects of these training activities would likely be subtle. Therefore, utterances elicited by an interview were transcribed and the accuracy, fluency and complexity were quantified in the way found in many previous studies. The students showed significant improvement in complexity, but any improvement in accuracy was found. They used less fillers, self-corrections, and repetitions, but instead, they paused longer and more frequently.
Principal Investigator: Takafumi Hayashi
Title: An Intelligent for the Integration of Sensor Data and Related Services for Healthcare, Environmental monitoring, and smart-grid management
Amount: ¥1,888,000
A new approach to the construction of a kind of intelligent infrastructure providing loosely coupled integration of sensor data and related services was presented. The proposed scheme uses a content-aware messaging network which can be constructed as a structured overlay network. Another novel approach of the primary and secondary data store grids for sensor network was also presented. The data store grids helps to construct and manage as secure, flexible, elastic, and sustainable loosely coupled integration of sensor data and related services. The policy mediation for over-lay networks having each own policies enables secure overlay networks inter-operation. The proposed scheme can be realized by a structured overlay network over an IP network.
In the project the following were studied.
Blue-tooth based healthcare system based on the Continua Health Alliance standard
Sensor-data stores for health-care measurements were constructed using keyvalue data stores and cryptography techniques
The integration of the systems above mentioned with related web-services by using message mediation were studied
In a similar manner, a network infrastructure for the integration of weather information and related services will be enhanced.
Security of SOA-based system using physically random number generator and secure distribution technique.
Principal Investigator: Saji N. Hameed
Title: Aizu Weather-Development of Business Sensitive e-Weather Services for the Aizu Region
Amount: ¥988,000
Environmental information systems provide accurate and timely estimates of the state of the environment which surround us. Such timely information is useful not only for decision makers at various levels, but also for the common public in carrying out their day to day activities. Information flowing out of such a system should not only be accurate and timely, but also should be easily comprehensible and possibly also provide some information on the future evolution at various time scales.
The Aizu weather project will develop needed systems to provide fast and accurate environmental information and forecasts. Environmental information will be generated by deploying a wide array of sensors with the help of the local community and public/private organisations in various parts of Fukushima. Weather forecasts will be generated using a well known operational weather forecast model running on grid scales close to 1 kilometer or less. Appropriate visualization strategies will be developed to convey the weather information accurately and comprehensibly. A simple but intuitive web interface will be developed that allows the share and access of environmental information through the internet.
Principal Investigator: Qiangfu Zhao
Title: Card holder Authentication Based on Image Fusion
Amount: ¥1,330,000
The purpose of this research is to propose a new way for card holder authentication. In 2009, we proposed a morphing based method, and filed a patent. However, this method cannot be used easily because the required memory size is big. To solve this problem, last year, we proposed a new approach based on linear image fusion. In fact, any image can be approximated by linear combination of several basis images. If the images are clustered properly using the k-PCA algorithm proposed by us earlier, each image can be represented by a small number of coefficients. To reduce the number of coefficients further, we proposed morphing based non-linear image fusion. Using this method, any face image can be approximated well using 2 or 3 reference images, provided that they are warped first based on a common skeleton. We have filed 3 new patents based on the results obtained last year.
Principal Investigator: Toshiaki Miyazaki
Title: Die-hard sensor network for monitoring disaster-hit region
Amount: ¥1,729,000
In the case of disaster, quick and detail surveillance of the damaged area is very important to make a rescue plan to save human life. However, the detail surveillance is usually done by hands in actual rescue scenes because there are no effective aids. Our goal is to develop a wireless sensor network (WSN) system, so called !Fdie-hard sensor network!G, which can monitor the disaster-hit region automatically by only scattering many sensor nodes in the region. The WSN has the following advantages; (1) no individual setup action is required in the field, (2) whole WSN system is never down and continuously monitors the disaster-hit region even if some sensor nodes involved in the WSN die, and (3) it has so scalable that we can expand the monitoring area and improve the sensing quality by only deploying extra sensor nodes as many as needed. In addition to these features, this system has a good cost performance compared to other surveillance systems such as a remotecontrolled helicopter. Thus, it can be applied to many critical disaster-hit regions, and we believe that our system must contribute not only to the rescue processes, but also to protecting the rescuers from secondary disasters. As the third year of the project, we concentrated on enhancing or developing the following technologies to realize the target system. (a) Reliability of wireless communications, (b) Location estimation of the sensor nodes, (c) Terminal software, (d) Sensor node design, and (f) Function alternation mechanism.
Principal Investigator: Keitaro Naruse
Title: High-Energy-Efficiency Field-Mobile Robot System
Amount: ¥847,000
In this research project, we assume only a small number of accurate landmarks and a limited number of mobile and inaccurate landmarks are available, and we have developed a localization system with those landmarks. We called it as the swarm mutual localization system for mobile robots. We have derived the extended
Kalman filter for our sensor system, and investigated the error accumulation. The results showed that the developed system has capable for the practical use if it is in a designed range.
Principal Investigator: Michael Cohen
Title: Cybervehicles
Amount: ¥1,061,000
We explore the intersection of visual music, ubiquitous computing, and vehicular informatics. In this project, we focus on the display capabilities of cars, via windshield wipers and spatial sound. Inspired by the ubicomp (ubiquitous computing) vision in which all aspects of an environment have potential I/O capability, we investigate the ability of windshield wipers deployed as output devices, synchronized to music, and the potential of spatial sound in cars for both passengers and drivers.
Principal Investigator: Shinya Oku
Title: Color Visualization of Human Health Conditions: Algorithm Establishment and Implementation for a!HSpecified Health Counseling Program!I(Statutory Tokutei Hoken Shido)
Amount: ¥1,427,000
We developed method to visualize the patient's health status for specified health counseling system (tokutei-hokenshidou system) by analyzing open medical questionnaire and data from the insurer. We also implemented it as an iPad library to be used in application development, although this technology is not implemented into the application yet. We would like to enhance this technique and implement this for a better operation of the application in clinical setting. The color system seems to be able to display human health status. At least, displaying simultaneously the values of physical, mental, and social health, it seems to be easy to use by medical professionals and related general public.
Principal Investigator: Kenichi Kuroda
Title: An FPGA/GPU-base Desktop High Performance Computing
Amount: ¥1,946,000
Supercomputers, such as "Blue Gene"or"Earth Simulator", show quite high computing performance but they are extraordinary expensive. We are aiming low-cost high performance systems based on FPGA/GPU-based systems with low investment. (∼1/1000)
Achievements
Design and Implementation of a Portable Framework for PCI Express Interface The proposed framework provides DMA transfer and interruption between a host PC and an FPGA board to reduce development time for acceleration board implementation. The proposed framework provides an interface circuit designed by using Altera's SOPC Builder to enhance portability. The proposed framework functions on different Altera's FPGA boards to support PCI Express. The proposed framework is easy to be transplanted. A new PROGRAPE system, specialized for interparticle calculation, with the proposed framework is implemented on two different size FPGA boards to evaluate portability and the performance of interparticle calculation. As a result, the new PROGRAPE system is easy to be transplanted to a larger FPGA board. The transplanted system achieves higher performance without significant redesign.
GPU Acceleration of Smoothed Particle Hydrodynamics Simulation We utilized GPGPUs to accelerate the Smoothed Particle Hydrodynamics (SPH) calculation, one of the fluid dynamics calculation methods. When the particle exceeded 8,000 in 2D version, the speed become six times faster than a signle core CPU. The rise afterwords become moderate and saturated. When the particle exceeded 5,700 in 3D version, the speed became five times faster. This shows usefulness of GPGPU for high performance desktop supercomputing with low costs.
Principal Investigator: Ben A. Abderazek
Title: Embedded Wireless Sensor Network System for Elderly Health Monitoring
Amount: ¥869,000
Recent technological advances in wireless networking, microelectronics and the Internet allow us to fundamentally modernize and change the way health care services are deployed.
Thus, the anticipated change and new services are urgently needed to help cope with the imminent crisis in the health care systems caused by current demographic,
social, and economic trends in Japan and in most developed countries.
The ultimate goal of the project is to research and develop a prototype of an in-body wireless sensor network system to assist old people by monitoring their
health status remotely by introducing smart in-body sensor based system that allow medical professionals to monitor elderly health status. The novel system is based on a wireless micro-nodes where a node incorporate miniaturized sensor(s), and special processing element.
Principal Investigator: Naohito Nakasato
Title: Development of Computer Network Architecture and Parallel Algorithms with Local Interconnect
Amount: ¥2,852,000
In recent parallel computer architecture research, we face a several severe difficulties such as demand for lower power consumption and gap between speed of a processor and Input/Output (I/O) speed. The inter-processor I/O is important research area related to parallel computers that is our focus in this project. We have designed a special computer cluster with the local interconnect. For this purpose, we have developed a network interface board for the inter-processor I/O. Moreover, we have developed and implemented new parallel algorithms and parallel applications on many-core systems as GPUs and Cell BroadBand Engine.
2011 THE UNIVERSITY OF AIZU ALL RIGHTS RESERVED