ゼミナール発表

日時: 1月6日(水)3限 (13:30-15:00)


会場: L1

司会: 南 裕樹
藤山 卓己 1551093: M, 1回目発表 ロボティクス

Title: Development of electric prosthetic fingers with built-in actuators for aesthetic consideration.

Abstract: Finger amputees usually use prosthetic fingers. Cosmetic fingers resemble human fingers, but they are inconvenient for working because joints cannot move. Electric prosthetic fingers can move joints with actuators, but they are less aesthetic than cosmetic fingers. In this work, we develop electric prosthetic fingers which combine both aesthetic appeal, through a design similar to a human finger, and functionality, by making all the joints movable.

発表題目:装飾性を考慮したアクチュエータ内蔵型電動義指の開発

発表概要:指を切断した際,通常,義指が使用される.装飾義指は外観が自然であるが,関節が駆動しないため作業には不向きである.電動義指は関節をアクチュエータで駆動できるが,装飾義指に比べると外観が自然ではない.そこで,本研究では指に近い自然な外観とすべての指関節を駆動することができる,電動義指の開発を行う.

 
LAO BRYAN TSANG 1551127: M, 1回目発表 数理情報学
title: Development of Automated Sit-to-Stand Rehabilitation Tool
abstract: The sit-to-stand (STS) movement is a functional task that is an essential pre-requisite to most activities of daily living. Teaching proper STS motion through rehabilitation, the quality of life will be improved for the elderly population. STS rehabilitation usually involves long-term, one-on-one therapist-patient sessions which are repetitive and not always accessible. Thus, automating this process is an attractive option. In this study, we aim to determine key indices for good STS rehabilitation by analyzing the physical and physiological differences of motion induced by therapists of different skill levels. We aim to use these results to understand effective STS therapy and to develop an automated rehabilitation system.
language of the presentation: English
 
笠井 裕貴 1551031: M, 1回目発表 大規模システム管理
title:Automatic Evacuation Guiding Scheme Using Trajectory of Mobile Nodes in Congested Road Networks
abstract:When large scale disasters occur, evacuees have to evacuate to safe places quickly. In recent researches, an automatic evacuation guiding scheme using mobile nodes has been proposed by Komatsu et al. In the research, each mobile node tries to navigate its evacuee by presenting an evacuation route. However, the scheme hasn't been considered traffic congestion when people evacuates. In this study, we aim to propose a new evacuation guiding scheme in congested road networks.
language of the presentation: Japanese
 

会場: L2

司会: 藤本 まなと
藤本 啓輔 1551092: M, 1回目発表 コンピューティング・アーキテクチャ
title: Dynamic Power Estimation of FPGA Accelerator
abstract: In the future, the use of FPGA increases. In particular, it is used mainly in large-scale system such as supercomputers. As a problem, performance of the FPGA limited by power consumption. Therefore, it is necessary to improve power efficiency for performance improvement. So, the developers can use performance of FPGA to the maximum by knowing the power of FPGA. But, there is no general power measurement method. This study estimates dynamic power of FPGA accelerator. First, I estimate whole power of FPGA using performance counter. Next, I estimate partial power of FPGA. Eventually, I create a framework to lower design cost.
language of the presentation: Japanese
 
一ノ瀬 智浩 1551010: M, 1回目発表 ソフトウェア工学
title: Removing Self-Admitted Technical Debt with Gamification
abstract:
Technical debt is a metaphor of software development that has been used to express non-optimal solutions, for example, incomplete, requiring rework, producing errors, or a temporary workaround.
It can accelerate development in the short term. In contrast, it causes some problems in the long time. Self-admitted technical debt can be seen in source code comments with keywords such as "TODO" or "FIXME" in many OSS projects.
In this study, I develop a system that supports removing self-admitted technical debt with visualization and gamification.
The system visualizes the directory structure of the software as a city, makes self-admitted technical debt obvious, and shows the ranking of the number of removed self-admitted technical debt by each developer for motivating them to remove them.
language of the presentation:Japanese
 
荒川 周造 1551005: M, 1回目発表 ユビキタスコンピューティングシステム
title: The study of rent estimation system using the traffic sensor
abstract: The decision of real estate rate has been made based on the tacit knowledge that veteran salesman is made from cultivated intuition and experience, the conversion to explicit knowledge (expression) has become an issue. On the other hand, in recent years it has become possible the collection of any information due to improvements in sensing technology. Therefore, we construct evidence-based rent estimate system by sensing and mining wide variety of data using the physical sensor. In this study, we report that we have developed a sensing system for obtaining the sense of traffic volume.
language of the presentation: Japanese
発表題目: 通行量センサを用いた賃料推定システムに関する研究
発表概要: 不動産の賃料の決定は,ベテラン営業マンが培った勘や経験からなる暗黙知を元になされており,その形式知化(表出化)が課題となっている.一方で,近年ではセンシング技術の向上によって,あらゆる情報の収集が可能となっている.そこで物理センサを用いて多種多様なデータをセンシング・マイニングすることで,根拠に基づく賃料推定システムを構築する.本研究では,通行量のボリューム感を得るためのセンシングシステムを開発したので報告する.
 
奥 雅和 1551026: M, 1回目発表 情報基盤システム学

title: Problems of SDN and the introduction of a previous study

abstract: SDN has attracted in the respect of network virtualization. SDN separates into controller which controls a route and switch which controls a transmission. SDN has the advantages of programable and flexible configuration but it also has problems. The controller is a problem of SDN because it has important information such as network topology. Studies of SDN have been conducting, but the number of the studies which aim to solve the problem of controller is few. In this presentation, I will introduce my study which aims to solve the problem of SDN and a previous study.

language of the presentation: Japanese

 
弥谷 圭朗 1551111: M, 1回目発表 モバイルコンピューティング
title: Measurement of the leakage of information by the timing attack
abstract: A timing attack that guess a secret key by observing the running time of cryptosystem is a serious threat. Therefore, it is important to evaluate the risk of timing attack. the leakage of information can be evaluated by calcuating the entropy of multinominal distribution. We derive an approximate expression of the entropy of multinominal distribution.
language of the presentation : Japanese
 

会場: L3

司会: 進藤 裕之
西 諒一郎 1551071: M, 1回目発表 視覚情報メディア
title: Inner slice visualization using a pair of paraboloidal mirrors
abstract: This study proposes a new system that photographs inner slices of three-dimensional objects. The proposed technique uses an optical system that consists of a pair of paraboloidal mirrors whose vertex and focal point correspond with each other. We can capture a specific layer of an object by setting the object at a focal point of a paraboloidal mirror and an image sensor at the focal point of the other mirror because light rays from the focal point of the paraboloidal mirror gather at the focal point of the other one. This proposed system has two advantages. The installation of the camera and mirrors of our system is easier than those in previous works. We do not have to consider the light wavelength because this system uses mirrors instead of lenses. In experiments, we confirmed that the proposed system can photograph each layer of an object that has two layers. In addition, we measured the degree of blur in captured images for point light sources at different positions in simulation.
language of the presentation: Japanese
発表題目: 一対の放物面鏡を用いた断層撮像
発表概要: 本研究は,3次元物体の特定の層を撮影するシステムを提案する.提案システムでは、互いの頂点と焦点を一致するように設置した一対の放物面鏡を用いる.一対の放物面鏡は一方の鏡の焦点から出た光はもう一方の鏡の焦点へ集光するという特徴をもつため,物体を一方の焦点位置に設置し,もう一方の焦点位置に撮像素子を設置することで,物体の特定の層の撮影が可能となる.提案システムは,同様の目的を持つ従来研究に対して比較的カメラや鏡の取り付けが簡単で,また,レンズを使用しないため光の周波数を考慮する必要がないという特長を持つ.実験では,2層構造をもつ物体のそれぞれの層が撮影できることを確認した.加えて,シミュレーション環境で異なる位置に光源を設置し,撮影される画像のボケ度合の測定を行った.
 
伊佐 衣代 1551007: M, 1回目発表 知能コミュニケーション
title: Statistical speech modification method capable of controlling movements of vocal organs
abstract: A conventional speech modification framework based on direct modification of acoustic speech parameters, such as spectral envelope and fundamental frequency (F0), is hard to intuitively modify speech signals. To make it possible to more intuitively and flexibly control speech signals, we propose a speech modification method capable of controlling movements of vocal organs. Transformation between the acoustic speech parameters and the physical movement parameters is achieved by using statistical models. In this presentation, I explain the related conventional research and my proposed method and describe my future research plan.
language of the presentation: Japanese
 
吉川 将司 1551119: M, 1回目発表 自然言語処理学
title: Syntactic Analysis to Spontaneous Speech
abstract: There is a growing need to give a syntactic analysis(to parse) to automatic speech recognition(ASR) output, such as speech translation and intelligent personal assistant application, because this technology provides the disambiguated structure of the input sentence for these downstream applications. This research aims to foster the existing algorithm to parse disfluent speech input so that it even be robust to ASR misrecognition.
language of the presentation: Japanese
 
佐藤 大夢 1551049: M, 1回目発表 光メディアインタフェース
title: Improvement of Reality for CG Contents by Projecting around the Display
abstract: A lack of a photometric consistency between real scenes and CG contents interfere a realistic appearance of CG contents (in the real scenes).In our study, we propose a method using a projector to obtain the photometric consistency between real scenes and CG contents in order to show it more realistically. Therefore we project a shadow around the display showing the CG contents so that it appear to be illuminated by an imaginary light source. We confirmed the effectiveness of our method by comparing the appearance of presence/absence of the projection.
language of the presentation: Japanese
発表題目: ディスプレイ周辺環境への光投影による表示CGの現実感の向上
発表概要: ディスプレイを用いた実世界上でのリアルなCG表現において,ディスプレイの光源としての性質による実世界-表示CG間の光学的整合性の欠如は高い現実感を得ることの妨げとなる.そこで本研究では,プロジェクタを用いて実世界-表示CG間の光学的整合性を確保し,表示CGの現実感を向上させる手法を提案する.ディスプレイ周辺環境へ影を投影し架空の光源を想起させることで,ディスプレイを光源ではなく照らされているように見せる.投影前後でのCGの見え方を主観評価し,提案手法による見た目の違和感の低減効果を確認した.
 
魚谷 果那 1551017: M, 1回目発表 環境知能学
title:[Paper Introduction] A Linked Visualization of Trajectory and Flow Quantity to Support Analysis of People Flow
abstract: Thanks to the recent evolution of movie- and sensor- based human tracking technologies, we can obtain and accumulate a set of walking paths (trajectories in this paper) of people over a long period in various places. Such people flow datasets are useful for many fields, including analyses of customer behavior, effectiveness of advertisements, and operational efficiency. We presents a linked visualization system to assist in the discovery of new knowledge by analyzing the accumulated people flow datasets, and a case study using this system. In this study we suppose the people flow datasets consist of a set of trajectories and temporal flow quantity. The system consists of two visualization components: classified trajectory visualization, and temporal flow quantity visualization. The former component classifies trajectories into several patterns applying the spectral clustering algorithm, and visualizes the patterns by colors on a physical space. The latter component displays temporal flow quantity of the above patterns applying a piled polygonal chart. This paper introduces a case study applying a movie-based human tracking dataset to the presented system.
language of the presentation: Japanese
 
BROU WIILIAM-FABRICE 1551121: M, 1回目発表 ネットワークシステム学
title: Multiple-Input Multiple-Output Dynamic Charging System Using Parallel Line Feeders
abstract: Dynamic charging refers to the capacity for a device (usually a electric car) to be powered/charged while in motion using wireless power transfer technology which means no wire/cable contact between the device and the source of power.
The wireless power transfer technology has been used for many years in applications like computer mouse, toothbrush, smartphone, car parking where the receiver has to stay immobile on a charging station to receive the power.
Nowadays, the challenge is to be able to dynamically power/charge a moving receiver. Many researches try to adapt the aforementioned immobile techniques for mobile applications but extending their range makes them quite complicated and too expensive to build.
Instead, the wireless power transfer in this research uses parallel lines as feeder/transmitter which benefit from its long primary coil structure, is simple enough and promising for applications including the receiver's motion such as an electric car on the road.
But, despite its promising advantages, the parallel line feeding system suffers from issues and one of them is the severe fluctuation of the receiver's power profile due to the standing wave in the electric and magnetic fields existing all along the feeding line which leads to some points of the line where no power can be received.
This step of the research presents an experimental configuration using multiple parallel line feeders capable of overcoming the standing wave issue and improving the output received power.
language of the presentation: English