日時: 11月12日(月)3限 (13:30-15:00)

会場: L1

司会: 渡辺一帆
加藤 綾華 1251031: M, 1回目発表 数理情報学
title: Analysis on Gaze control in terms of driving experience for lane change
abstract: I analyze the difference of eye gaze between expert driver and non-expert driver. And I intend to establish the teaching way to improvement non-expert driver's skill. I conducted a pilot experiment with lane change task which requires driving skills. As a result, non-expert driver does not check the rear-view mirror too much. So, it was seen that non-expert driver could not keep inter-vehicle distance, they clashed in the worst case. I also introduce more reliable drivinng and eye gaze observing system.
language of the presentation: Japanese
発表題目: 運転頻度別でみた車線変更タスクにおける視線解析
発表概要: 熟練、非熟練ドライバーの視線を分析し、非熟練ドライバーの 運転技術向上を助ける享受法を確立することを目的とする。熟練ドライバー・非熟練ドライバーにおいて運転スキルで差が生じる車線変更タスクを用い、パイ ロット実験を行った。その結果として、運転行動時に非熟練ドライバーでは特 にバックミラーの確認が少なく、そのために合流後に車間距離がうまく 取れず最悪の場合衝突する、などの違いが見られた。また、現在作成中のより信頼度の高い運転 視線観測システムについても紹介する。
春日 照之 1251030: M, 1回目発表 ロボティクス
title: Evaluation Method of Artificial Ankle Joint Function
abstract: It is important for Total Ankle Arthroplasty to evaluate the artificial ankle joint function which is very important for walking. In this presentation, the purpose of this research and related works are shown. I will introduce accuracy of X-ray image in determining relative position and orientation of total knee replacement components. Then I will describe arch analysis of the human using motion capture and force plate.
language of the presentation: Japanese
柳井 一成 1251109: M, 1回目発表 生命機能計測学
title: Study on the development of body temperature monitoring system to support the proactive behavior of spinal cord injuries (SCI)
abstract: SCI is an instrument thermoreceptor and thermoregulatory effect does not appear to operate normally, there is a risk of developing Hypothermia under cold environment or depression under hot environment. Therefore, when they go out they need to know in real time the temperature in particular. So, we aimed to develop devised a monitoring system for an SCI temperature in this study. In this presentation, I describe research background and relevant research.
language of the presentation: Japanese (choose one)
発表題目: 脊髄損傷者(脊損者)の主体的行動を支援するための体温モニタリングシステムの開発に関する研究
金川 雅和 1251032: M, 1回目発表 知能システム制御
title:Stable power supply on the electric power system under electricity liberalization
abstract: In recent years, Electricity liberalization is in the spotlight. For stable power supply, all providers in power system need to cooperate to hold same frequency and same voltage phase angle.But all providers may not cooperate under electricity liberalization.KKT Price-Based Control was proposed for such problem.In this presentation,I will introduce the KKT Price-Based Control and its demerit.
language of the presentation: Japanese
発表題目: 電力自由化が導入された電力系統における電力の安定供給
発表概要:近年, 日本において電力自由化が注目を集めている.電力の安定供給を実現するためには, 電力系統内の発電会社が協調して電圧の周波数や位相を一定にしなければならない.しかし電力自由化には, 電力系統内の事故などによって周波数や位相が目標値からずれてしまった場合, 発電会社が協調して周波数や位相を一定にするとは限らないという問題がある.先の問題対してKKT Price-Based Controlと呼ばれる手法が提案されている.本発表ではKKT Price-Based Controlとその欠点について述べる.

会場: L2

司会: 奥田 剛
上田 知幸 1251015: M, 1回目発表 ソフトウェア基礎学

title: An efficient travel scheduling for Electric Vehicles based on Car-Sharing service

abstract: Electric vehicles (EV) have become popular in the past several years. However, the short travel distance and the long charge time are big problems. In this study, we plan to adapt EVs to long distance travel by using car-sharing service. The idea is allowing driver change his/her EV to other full charged EV before the current EV run out of battery. The target of our proposed method is to find an optimal car-sharing schedule in a multiple drivers/EVs environment.

language of the presentation: Japanese

発表題目: カーシェアリングに基づいたEVの効率的運用スケジューリングの検討

発表概要: 近年,電気自動車(EVElectric Vehicle)は徐々に普及するようになったが,EVの短い航続距離と長い充電時間はさらなる普及に向けた課題となっている.本研究では,カーシェアリングの仕組みを利用しユーザが電力残量が少ないEVを他のEVに乗り換えることで長距離走行を実現する.提案手法では,多数のカーシェアリング利用者を対象に,EV乗換スケジューリングを行う.

楠田 浩平 1251039: M, 1回目発表 コンピューティング・アーキテクチャ
title: Paper introduction "CPU-Assisted GPGPU on Fused CPU-GPU Architectures"
abstract: This paper presents a novel approach to utilize the CPU resource to facilitate the execution of GPGPU programs on fused CPU-GPU architectures. In our model of fused architectures, the GPU and the CPU are integrated on the same die and share the on-chip L3 cache and off-chip memory, similar to the latest Intel Sandy Bridge and AMD accelerated processing unit (APU) platforms. In our proposed CPU-assisted GPGPU, after the CPU launches a GPU program, it executes a pre-execution program, which is generated automatically from the GPU kernel using our proposed compiler algorithms and contains memory access instructions of the GPU kernel for multiple threadblocks. The CPU pre-execution program runs ahead of GPU threads because (1) the CPU pre-execution thread only contains memory fetch instructions from GPU kernels and not floating-point computations, and (2) the CPU runs at higher frequencies and exploits higher degrees of instruction-level parallelism than GPU scalar cores. We also leverage the prefetcher at the L2-cache on the CPU side to increase the memory traffic from CPU. As a result, the memory accesses of GPU threads hit in the L3 cache and their latency can be drastically reduced. Since our pre-execution is directly controlled by user-level applications, it enjoys both high accuracy and flexibility. Our experiments on a set of benchmarks show that our proposed preexecution improves the performance by up to 113% and 21.4% on average.
language of the presentation: Japanese
発表題目: GPU搭載型CPUにおけるCPUアシストGPGPU技術(論文紹介)
発表概要: 本論文では、GPU搭載型CPUにおいてGPGPUプログラムの実行を改善にするための新たな手法を提案する。 我々の仮定するモデルとして昨今のIntel社製SandyBridgeコアやAMD社製のAPUなどが挙げられ、これらは同一ダイ上にGPUとCPUを搭載し、L3キャッシュやメインメモリを共有している。 提案するCPUアシストGPUシステムは、CPUで事前処理されたプログラムとGPU用のプログラムを並行して実行していく。 GPUプログラムがスタートした後、CPUは前処理されたプログラムを実行する。 この前処理されたプログラムは、新たに提案するコンパイラアルゴリズムによってGPUマルチスレッド用のメモリアクセス命令を利用して GPUカーネルを基に自動的に生成される。 CPU上の前処理されたプログラムのスレッドは、GPUスレッドに先行して実行される形となる。 理由として、 (1) 前処理されたプログラムはGPUカーネルからのメモリフェッチ命令のみを含んでおり、かつ浮動小数点/整数演算をサポートしないから (2) CPUは高い周波数で動作し、GPUのスカラコアよりも高度な命令レベル並列化を行うことが出来るから などが挙げられる。 また、L2キャッシュにおけるプリフェッチャを強化することによりGPUスレッドからのメモリアクセスがL3キャッシュでヒットし、メモリアクセス時のレイテンシを大幅に低減した。 これらの技術がユーザレベルアプリケーションで直接コントロールできるようになれば、効果に柔軟性と精度が増すと考える。結果的に、この前処理技術によって最高で113%、平均して21.4%のパフォーマンス改善が見られた。
尾上 佳久 1251027: 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: A Wi-Fi Offloading Method Based on Prefetching
abstract: Mobile data traffic is growing at a drastic rate. This growth is due to the spreading of the use of smartphones, the increasing number of subscribers, and the rapid development of mobile broadband. This explosive growth of the number of mobile users causes immense pressure on the limited spectrum of cellular networks (3G networks), and the deterioration in the quality of wireless communication. To cope with this problem, mobile network operators deploy Wi-Fi access points(Wi-Fi APs) to offload the traffic from 3G networks to Wi-Fi. However, since the coverage of a Wi-Fi AP is relatively small and the number of Wi-Fi APs are limited, Wi-Fi connectivity is far from ubiquitous. There are some existing offloading methods that postpone communication requests from applications in 3G networks until the user moves to a Wi-Fi area in the near future. These methods, however, cause the poor performance of delay sensitive applications like web browsing and streaming of video and audio. To address this problem,we propose a Wi-Fi offloading method based on prefetching. Our method predicts the contents that will be requested by the user in the 3G networks, and prefetches them in advance when Wi-Fi connectivity is available. Thus if the prefetched contents are requested after the user moves from the Wi-Fi area to a 3G area, the user can instantly access the contents without communication over 3G networks. In this presentation, I will talk about the background of our research, related studies, and our proposed method.After that, I will show what has been done so far, and present my research plan.
language of the presentation:Japanese
発表概要:スマートフォンの普及,携帯電話利用者の増加,モバイル・ブロードバンドの普及によりモバイル・データ・トラフィックが急増している.これにより,携帯電話回線(3G回線)の負荷が高まり,それに伴う通信品質の低下が危惧されている. このような状況から,3G回線の負荷をWi-Fiネットワーク側に逃がすWi-Fiオフローディングが注目されている.しかし,1つのWi-Fiアクセスポイント(Wi-Fi AP)のカバレッジは狭く,またその設置台数も少ないため,Wi-Fiオフローディングの機会が少ないという問題がある. この問題に対して,既存研究では,3Gエリアで発生した通信要求を,近い将来Wi-Fiエリアに入るまで遅延させることで,オフローディングの機会を拡大する方法を提案するものがある. しかしこの方法は,高い応答性が求められるコンテンツの閲覧に対しては適用が困難である.この問題を解決するために,本研究では,プリフェッチに基づいたWi-Fiオフローディング手法を提案する. 提案手法では,ユーザがWi-Fiエリアに滞在中に,そのエリアを出た後に3Gエリアで閲覧する可能性があるコンテンツを予測し,事前にダウンロード(プリフェッチ)する.これにより,ユーザが3Gエリアに移動した後,プリフェッチ済みのコンテンツが要求された場合に,3G回線を利用せずに即座にコンテンツの提示が可能となる. スマートフォンにとってバッテリは最重要のリソースであるため,この手法を提案するにあたって消費電力に対する制約を設ける.そして制約では,あるコンテンツをプリフェッチと,プリフェッチできなかった文を3G回線で取得した場合の消費電力量の合計が,同じコンテンツを3G回線のみで取得した場合の消費電力量を超えないようにする. また提案手法では,プリフェッチするコンテンツを,その容量と閲覧確率より導き出したオフローディング期待を元に評価する.そして,オフローディング期待値が最大になるようなコンテンツの集合をプリフェッチすることによって,オフローディング量を最大化する. 本発表では,研究背景,関連研究,提案手法,および進捗と今後の課題について述べる.
青江 丈 1251001: M, 1回目発表 インターネット工学

Title:Improvement of underwater network simulator


Underwater networking has many applications and protocols are proposed

To test new protocols, simulation or field test are necessary

Simulation has 4 advantages compared with field test

Underwater acoustic networking has many difficulties so we cannot simply use terrestrial wireless simulators

Previous underwater networking simulators have 3 problems

I want to create simulator modules which overcome these problems based on NS3

Language of the presentation: English


会場: L3

司会: Graham Neubig
吉本 一平 1251119: M, 1回目発表 自然言語処理学
title: Error detection using deep learning
abstract: We introduce the basic idea of deep learning, an emerging field of machine learning, and its application to error detection of sentences writen by the English learners. The main advantage of deep learning is its ability of learning multiple levels of representation and abstraction that make sense of data.We utilize this ability to obtain good features for the task of error detection for language learners in which the performance is greatly affected by feature selection.
language of the presentation: Japanese
石堂 なつみ 1251008: M, 1回目発表 音情報処理学
title: An auditory ERP-based BMI for a universal controller in a real environment
abstract:Recent years, there are a lot of researches about Brain-Machine Interface (BMI) that is a direct communication pathway between the brain and an external device, and BMI realize to control electrical appliances without using hands or any other body parts. However, past experiments were only held in experimental room, and we seek to realize it in real environment using "BMI house". In this presentation, I will suggest the BMI especially based on auditory attention, and show some experimental results.
language of the presentation: Japanese(power point is in English)
吉田 理貴 1251117: M, 1回目発表 知能コミュニケーション
title: Group Dialogue Support by use of Statistical Dialogue Models
abstract: The purpose of this research is to investigate how system supports group dialogue. First of all, for this purpose we have to inquire into Dialogue Model. The reason why we researched this model is that it gives the most appropriate action for system. This is the central problem of the dialogue action. So, in this presentation the main stress falls on the investigation of the dialogue model.
language of the presentation: Japanese
石黒 景亮 1251007: M, 1回目発表 環境知能学
title: The Effect of Electronic Self-Monitoring on Weight Loss and Dietary Intake: a Randomized Behavioral Weight Loss Trial
abstract:Technology may improve self-monitoring adherence and dietary changes in weight loss treatment. Our study aimed to investigate whether using a personal digital assistant (PDA) with dietary and exercise software, with and without a feedback message, compared to using a paper diary/record (PR), results in greater weight loss and improved self-monitoring adherence. Healthy adults (N = 210) with a mean BMI of 34.01 kg/m2 were randomized to one of three self-monitoring approaches: PR (n = 72), PDA with self-monitoring software (n = 68), or PDA with self-monitoring software and daily feedback messages (PDA+FB, n = 70). All participants received standard behavioral treatment. Self-monitoring adherence and change in body weight, waist circumference, and diet were assessed at 6 months; retention was 91%. All participants had a significant weight loss (P < 0.01) but weight loss did not differ among groups. A higher proportion of PDA+FB participants (63%) achieved >= 5% weight loss in comparison to the PR group (46%) (P < 0.05) and PDA group (49%) (P = 0.09). Median percent self-monitoring adherence over the 6 months was higher in the PDA groups (PDA 80%; PDA+FB 90%) than in the PR group (55%) (P < 0.01). Waist circumference decreased more in the PDA groups than the PR group (P = 0.02). Similarly, the PDA groups reduced energy and saturated fat intake more than the PR group (P < 0.05). Self-monitoring adherence was greater in the PDA groups with the greatest weight change observed in the PDA+FB group.
language of the presentation: Japanese
ANGIE CHEN 1251122: M, 1回目発表 インタラクティブメディア設計学
title: An Augmented Reality application based on Experiential Learning Theory with better Observation for Improving Learning
abstract: Experiential learning is the process of creating knowledge from the transformation of experience. It is known as one of the most effective ways of learning. However in the traditional school learning environment, students are often limited to learn facts and principles by memorizing without observations and understandings. This research aims to improve students’ learning in daily lives experiences and in school. With the learning theory as foundation and the positive impacts for education that Augmented Reality (AR) has, the applicant proposes an educational AR application with tangible interface to improve students’ learning. The application provides an opportunity to students to observe the phenomena that are unobservable in reality.
language of the presentation: English