| 袴田 有哉 | 1451083: M, 1回目発表 | ロボティクス |
|
title: Stable motion of a humanoid robot in consideration of the contact force in manipulation
abstract:A humanoid robot is demanded to perform tasks on behalf of humans; it is advantageous that a humanoid robot has human-like figures. However, the robot must manipulate the various objects such as a desk and a door during the tasks and contacting force in manipulation possibly causes unbalance of the robot. The zero moment point(ZMP) is widely used as a stability index of the humanoid robot, but the ZMP does not necessarily works especially in the task where the robot applies force to the environment by the robot’s hand. In this presentation, I introduce the Capture point which is the technique to avoid falling down, and show the implementation in the robot’s simulation. language of the presentation: Japanese 発表題目: ヒューマノイドロボットによる作業中の接触力を考慮した安定動作 発表概要:人間に近い形を持つというメリットを生かし、ヒューマノイドロボットに人間の代わりに作業を行わせたいという欲求が高まっている．しかし，ロボットが作業を行う為には机や扉など様々な物に触れる必要があり，そこに生ずる接触力によって転倒する危険性がある．ロボットの安定指標としてはzero moment point(ZMP)が広く用いられているが，外力を発生する手作業においては使用できない場合が存在する．本発表では，転倒回避の手法の一つとしてCapture pointを紹介し，シミュレータ上で実装した例を示す． | ||
| 松永 拓也 | 1451099: M, 1回目発表 | ロボティクス |
|
title: Development of an assistive hand in the daily life for the person with impairment of upper limbs.
abstract: The paralysis of arm due to stroke reduces the quality of the daily life after the discharge. Therefore the exoskeleton hand for rehabilitation is developed actively, but it doesn't suppose using in the daily life. In this presentation, I suggest a new hand that can use functionally in the daily life for effective rehabilitation, and speak a future design policy. language of the presentation: Japanese 発表題目: 上肢機能障害者のための日常生活支援ハンドの開発 発表概要: 脳卒中による上肢の麻痺は退院後の日常生活の質を低下させる．これに対し，リハビリテーションを目的とした外骨格ハンドの開発が盛んに行われているが，日常生活での使用は想定されていない．本発表では，日常生活で機能的に使用できリハビリテーションに効果的な新しいハンドを提案し，今後の設計方針について述べる． | ||
| 江口 遼平 | 1451019: M, 1回目発表 | 計算システムズ生物学 |
|
title: Unraveling the defense mechanism of Angelica acutiloba by gene expression profiling
abstract: Angelica acutiloba is a perennial herb from the family Apiaceae or Umbelliferous. Its root was used in Kampo medicine, which is a Japanese adaptation of Traditional Chinese medicine. Angelica acutiloba has the defense mechanism to insect damage, and makes BETA-CARYOPHYLLENE which has the property of inviting a bee to exterminate noxious insects. In addition, BETA-CARYOPHYLLENE is a medically effective ingredient. In a word, medically effective ingredients in Angelica acutiloba increase by insect damage. The purpose of this study is unraveling the defense mechanism of Angelica acutiloba. In this presentation, I will report the analysis results so far. language of the presentation: Japanese | ||
| 岩根 史明 | 1451015: M, 1回目発表 | 数理情報学 |
|
title: Review of Co-Adaptive BMI system towards the application of Rehabilitation
abstract: Recent studies have shown that rehabilitation using brain-machine interface has a great potential in stroke patients with severe motor paralysis. The aim of this study is to do rehabilitation of foot motor ability using a co-adaptive brain machine interface. In this presentation, I am going to introduce a batch algorithm used in run throw demo and an online Co-adaptive algorithm.
language of the presentation: Japanese | ||
| 山崎 翔太 | 1451115: M, 1回目発表 | 知能システム制御 |
|
title: Vibration Control for Flexible Arm by Feedback Error Learning
abstract: Feedback Error Learning (FEL) is one of the adaptive control methods that adjust the parameters automatically. One of the previous works on the FEL is hand position control. However, the vibration on the hand position is not negligible because of weight reduction in recent years. Normally the vibration suppression is performed by feedback control, there is a problem about the fast response. Therefore, in this research, we consider the vibration suppression and hand position control of the Serial Flexible Link using FEL. In this presentation, we introduce the constitution of the FEL and describe a simple simulation result. language of the presentation: Japanese 発表題目: フィードバック誤差学習によるフレキシブルアームの振動制御 発表概要: フィードバック誤差学習(FEL)は, フィードフォワード制御器が逐次的に調整されることにより, プラント出力を目標値に追従させる制御手法である. FELを用いた先行研究の1つに, 剛体ロボットアームの手先位置制御があるが, 近年では産業用ロボットにおいても軽量化が望まれており, 手先位置に生じる振動が無視できなくなってきている. 従来, 振動抑制はフィードバック制御により行われるが, フィードバック制御では速応性向上が難しいという問題点がある. そこで本研究では, 制御対象として2リンク柔軟アームを取り上げ, FELによる振動・運動制御を実現することを目的とする. 本発表では事前確認としてFELの構成方法について確認し, 簡単な数値シミュレーション結果を示す. | ||
| 竹内 昌平 | 1451067: M, 1回目発表 | コンピューティング・アーキテクチャ |
|
title: A Sutudy on a Comprehensive Architecture Exploration Environment for Emerging Applications
abstract: Emerging applications, such as graph processing and machine learning, contain several irregular memory access patterns and control flows. While heterogeneous computing systems which consist of general CPUs and accelerators, such as GPU, have been widely used, the performance contribution given by each microarchitectual component of such hardware is still inapparent. In this paper, we propose a comprehensive processor evaluation environment that can evaluate various architecture settings. We evaluate the performance of triangle counting to identify the appropriate instruction window size for a single OoO core. language of the presentation: *** Japanese 発表題目: 次世代アプリケーションのための包括的な アーキテクチャ探索環境の検討 発表概要: グラフ処理や機械学習といった今後より重要になると考えられるアプリケーションの多くは，不規則な制御フローやメモリアクセスパターンを持つ．現在，汎用CPUに加えてGPUなどのアクセラレータを持つヘテロジニアスな計算機システムが普及しつつあるが，それぞれが持つ各ハードウェア要素がどの程度性能に寄与しているかは明らかになっていない．本研究では，次世代アプリケーションに適した効率的なプロセッサアーキテクチャ構成を模索するための，異なる構成を持つプロセッサを包括的に評価できるシミュレーション環境を実現する．また，初期実装により単一コア，単一スレッドでグラフ処理の1つであるトライアングルカウンティングの処理に有効な命令ウィンドウサイズの評価を行った． | ||
| 小河 亮 | 1451031: M, 1回目発表 | ディペンダブルシステム学 |
|
title: Burn-In Time Reduction Based on Test Data Analysis
abstract: Burn-In test is an effective test process to screen out early-life failures of VLSI circuits. However, the significant costs for equipments and long test time of Burn-In test are serious problems for manufacturers. Test data analysis is one of promising approaches to reduce these costs. Burn-In time can be reduced by predicting results of Burn-In with prediction model constructed from previously collected test data. In my presentation, I'm going to introduce an overview of related papers on test data analysis and my research progress. language of the presentation: Japanese | ||
| 水谷 早苗 | 1451102: M, 1回目発表 | ディペンダブルシステム学 |
|
title: Improved Power Supply Noise Control for Pseudo Functional Test
abstract: With shrinking of feature size of VLSI circuits, delay testing has become essential in order to ensure high product quality. However, delay testing typically consumes much higher power than functional operation. Excessive power dissipation during delay test may lead to timing failures due to power supply noise (PSN) induced delay increase, and thus it may result in undue yield loss (over-testing). Pseudo functional K longest path per gate (PKLPG) test is effective to make power during test closer to that of during functional operation, but should be carefully applied in order not to cause too low power dissipation condition that may cause under-testing. In this presentation, a novel method to generate PKLPG test patterns is presented. The proposed method intentionally maximizes switching activity around sensitized paths to avoid under-testing, while also keeping low power dissipation in the whole circuit to avoid over-testing. language of the presentation: Japanese | ||
| 前中 省吾 | 1451095: 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: On walking route recommendation through heart rate prediction
abstract: In recent years, walking is gathering attention as a preventive and improvement measures of adult diseases such as hypertension and diabetes diseases resulting from lack of exercise. However, high-perceived exertion during walking is a large burden on the heart and joints, and may decrease motivation for walking. Thus, for effective and sustainable walking, it is necessary to go for a walk with physical load suitable for individual physical ability.In this presentation, using the sensor data which can be acquired by smartphone, we propose a method for predicting heart rate change when walking along the route before walking and recommending the optimal walking route based on the predicted results. language of the presentation: Japanese 発表題目: 心拍数予測に基づく歩行経路推薦法に関する検討 発表概要: ウォーキングは高血圧，糖尿病等の運動不足に起因する生活習慣病の予防・改善対策として注目されている．しかし，負担の高い無理なウォーキングを続けた場合は，心臓や関節に大きな負担がかかり，ウォーキングに対する意欲の低下や継続性の低下につながる．したがって，継続性の高い効果的なウォーキングのためには，個人の身体条件に応じた適切な負担度でウォーキングを行う必要がある．本発表では，スマートフォンで取得可能なセンサデータに基き，任意の経路を歩行する際の心拍数変化を予測する方法と予測結果に基づき最適な歩行経路を推薦する方法を提案する． | ||
| 森下 慈也 | 1451112: 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: Support system for bicycle touring
abstract: Recently, we are needed clean and efficient moving method in the city from global environment problems and energy problem. Bicycle is gathering attention for alternative moving method from the car. Therefore, we are aimed to increase cyclist through hobby touring from usage of the bicycle to diversify. Bicycle touring is possible to run comfortably long distances if taking with someone, but it is difficult to find a bicycle rider who has a desired driving skills. In this presentation, modeling each user's skill and physical strength by using the sensor data which can be collected by smartphone, we propose a system that supports bicycle touring in several people who were gathered around a spot by modeled data. language of the presentation: Japanese 発表題目: 自転車運転者向けツーリング支援システム 発表概要: 近年のエネルギー問題や地球環境問題からクリーンかつエネルギー効率の高い持続可能な都市内交通体系の実現が必要となっており，自転車は自動車の代替移動手段として注目されている．そこで，自転車の利用形態が多様化していることから，趣味活動である自転車ツーリングを通して自転車人口を増加させることが狙いとなる．自転車ツーリングは複数人で行うと長い距離を楽に走ることができるが，ツーリングしようと思い立ったタイミングで運転技術の近い自転車運転者を探すことは難しい．そこで，本発表では，スマートフォンで取得可能なセンサデータに基づいて自転車運転者の運転技術をモデル化し，技術・体力が類似した走行中の自転車運転者を合流させ複数人でツーリングを行えるよう支援するシステムを提案する． | ||
| 枇榔 晃裕 | 1451088: M, 1回目発表 | インタラクティブメディア設計学 |
|
title: Tourist Information on deformed map with current place
abstract: During the last decade, electronic maps on mobile phones have become commonplace. Electronic maps, however, do not replace deformed maps that distributed on tourist spots or posted up on a board. In this study we focus on the characteristics of the deformation map that it is possible to look at the tourism point in list basis.The check and select tourism point was carried out on a deformation map, adds the location information on the deformation map you have made a selection of sightseeing point, and the guidance to the tourism point that you selected.We confirm the utility by performing seamless action from the selection point. In addition, we confirm the hypothesis that deformation map adding your location information is more useful than the normal electronic map based on general views adding your location information.We are discussed a method of adding a location information to the deformation map automatically. language of the presentation: Japanese 発表題目: 外部デフォルメマップに現在地情報を付加した観光地案内 発表概要: 携帯電話の進歩とともに、電子マップも発展し利用されるようになってきた。しかし、観光地で配布されたり、案内板で掲示されたりするようなデフォルメマップは携帯電話に搭載される電子マップに置き換わらずにいる。 本研究では観光地点を一覧的に眺めることができるというデフォルメマップの特性に着目し、観光する地点の確認や選択をデフォルメマップ上で行い、選択した観光地点までの案内を観光地点の選択を行ったデフォルメマップ上に現在地情報を付加することで、観光する地点の選択から行動をシームレスに行うことによる有用性を確認する。 そして、現在地情報を付加したデフォルメマップが、一般図を基とし現在地情報を付加した通常の電子マップよりも有用であるという仮説を確認し、デフォルメマップへ自動的に現在地情報を付加する手法を考察する。 | ||
| HUSEJIC AMAR | 1451123: M, 1回目発表 | インタラクティブメディア設計学 |
|
Title: Automatic Segmentation of Lumbar Spine CT images Abstract: Lower back pain is one of the most common medical problems and according to statistical information from around the world it will affect 8 out of 10 people at some point during their lives. In order to diagnose the problem doctors use X-ray and CT images. Comparing these images is hard for doctors, because X-ray images can be taken in any pose (painful pose) but they are in 2D and offer less detail compared to the CT image data which are in 3D and offer more detailed information but can only be taken in one pose (lying down). The system that is being developed in our laboratory solves this problem by modifying 3D CT image data to the pose of the 2D X-ray image. In order to prepare the data for the system we need to segment the spine bones which is a slow and tedious work when done manually. In order to solve this problem we are currently checking out existing methods so that we can come up with our own approach. | ||
| 西村 祥吾 | 1451081: M, 1回目発表 | 環境知能学 |
|
title: [Paper Introduction] APNEA Detection on Smart Phone
abstract: Statistics show that over 18 million Americans suffer from sleep apnea. In order to diagnose this syndrome, you must use polysomnography which is high restrictive detecting equipment. I introduce the article that suggests the method to distinguish the syndrome in real time by a system implemented on a smart phone. Moreover, I explain about the necessary vital data and the method of feature extraction features or classification for the diagnosis of apnea while the introduction of related work. language of the presentation: Japanese 発表題目: スマートフォン上での睡眠時無呼吸症候群検出 発表概要: 睡眠時無呼吸症候群によってアメリカ人だけでも1800万人以上が苦しんでいる。この症状を診断するためには、拘束性の高いポリソムノグラフィーを使用しなければならない。この種の患者の診断を、スマートフォン上で実装されるシステムによって無呼吸状態をリアルタイムに判別する方法を述べた論文を紹介する。加えて、SAHSの診断のために必要な生体情報および特徴料の抽出、分類手法について、関連研究に触れながら体系的に述べる。 | ||
| 松田 昇悟 | 1451098: M, 1回目発表 | 自然言語処理学 |
|
title: Coreference Resolution using Neural Network
abstract: Coreference resolution is the task of clustering a noun phrases that refer to same real-world entity in the text. In addition to improvement of the algorithm in the previous research, examination of the appropriate feature is performed. In this study, we examine the method of coreference resolution using word2vec. language of the presentation: Japanese 発表題目: ニューラルネットワークを用いた共参照解析 発表概要: 共参照解析は、文書中の名詞句を現実世界に存在する実体ごとに分類するタスクである。先行研究ではアルゴリズムの改善に加えて、適切な特徴量についての検討が行われている。本研究では共参照解析の特徴量としてword2vecの出力を用いる手法について検討する。 | ||
| 田尻 祐介 | 1451069: M, 1回目発表 | 知能コミュニケーション |
|
title: Noise Robust Non-Audible Murmur Enhancement Method Using Air-Conductive Microphone
abstract: Silent speech communication technology have recently been studied to enable speech communication without necessity of emitting an audible acoustic signal. Non-Audible Murmur (NAM) microphone can detect an extremely soft whispered voice and it is effectively used as silent speech interfaces. Moreover, it has greater usability than other devices such as EMG and ultrasound systems. On the other hand, quality and intelligibility of the detected voices severely degrade by body-conductive recording. To address this issue, NAM enhancement method based on statistical voice conversion is proposed. However, influence of external noise is not considered in the conventional method. Therefore, we propose a noise robust NAM enhancement method using air- and body-conductive microphones. language of the presentation: Japanese | ||
| 山根 壮一 | 1451118: M, 1回目発表 | 知能コミュニケーション |
|
title: The method of classification and control for singing voice for music creation support based on the expression word
abstract: We can post creation anyone freely on the internet by the recently grown of CGM (Customer Generated Media) culture. Above all, music activities including the song has been activity, is also expected to increasingly development in the future. However, there is a limit to vocal expression, we are not easy to represent any singing method or voice quality. In this study, I classified the singing voice based on the intuitive expression word, and propose to the method that anyone can control vocal expression easily. language of the presentation: Japanese 発表題目: 表現語に基づく歌唱音楽創作支援のための歌声分類制御手法 発表概要: 昨今のCGM(Customer Generated Media)文化の発展により， インターネット上に誰でも気軽に創作物を投稿できるようになった． 中でも，歌を含む音楽活動は活発に行われており，今後もますます発展が期待される． しかし，人間の歌唱表現には限界があり，自由に歌い方や声質を表現することは簡単ではない． 本研究では，歌声を直感的な表現語に基づいて分類し，誰でも簡単に歌声を制御できる手法を提案する． | ||