| LI SHANGLIN | M, 1回目発表 | 数理情報学 | 池田 和司, 川鍋 一晃（客員教授）, 杉本 徳和（客員准教授）, 田中 沙織 |
| 小澤 海斗 | M, 1回目発表 | 数理情報学 | 池田 和司, 荒牧 英治, 久保 孝富, 日永田 智絵 |
|
language of the presentation: Japanese
発表題目: Zero/few-shot classification of dog emotions from images 発表概要: abstract: It is often said that dogs and humans share a unique bond. That relationship spans thousands of years. Dogs are known to be loyal and affectionate, and are an essential part of the family for many people. Gaining a deeper understanding of dogs emotions not only enhances the bonds we share with our dogs but also enriches our collective lives by promoting mutual respect, understanding, and empathy. Fur- thermore, our ability to recognize and correctly in- terpret dogs emotional states is pivotal for effective communication, ultimately reducing potential con- flicts and fostering harmonious cohabitation. This is why research on dogs emotions is underway This paper aims to assess emotions using Dog images. To do so, We decided to use CLIP, one of the image classification models. | |||
| 洲澤 春樹 | M, 1回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 荒牧 英治, 諏訪 博彦, 松田 裕貴 |
|
title:
Proposal for A Coaching System Leveraging Generative AI for Adaptive Smartphone App Access Control
abstract: Smartphone applications such as social media and mobile games, create challenges for self-management by constantly distracting individuals and making it harder to focus on their personal goals. While many smartphones have built-in functions to limit application usage, this approach lacks the ability to effectively manage user motivation and offer adaptable coaching based on their efforts toward goals. This study proposes a system where Generative AI helps users make decisions about accessing smartphone applications based on their current situation, such as progress toward their goals and effort level. Currently, we are developing an iOS application where Generative AI allows users to use specific smartphone applications based on the progress in the English conversation training function within the application. language of the presentation: Japanese | |||
| 平良 繁幸 | M, 1回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 荒牧 英治, 諏訪 博彦, 松田 裕貴 |
|
title:Estimation of Response Attitudes Using Device Operation Information to Improve Crowd-Sourcing Response Behavior
abstract:Crowdsourcing enables large-scale data collection at low cost. However, the presence of respondents who prioritize task completion over accurate responses is a factor that reduces the reliability of survey results. In this study, we estimate response attitudes using machine learning based on users' device operation information. language of the presentation: Japanese | |||
| 笹崎 海利 | M, 1回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 荒牧 英治, 諏訪 博彦, 松井 智一 |
|
title: Examination of Distributed Representation of Microbehavior Through Speech
abstract: Microbehavior recognition is important for understanding behavior, but it is difficult to cover all the classes of behavior to be recognized in advance, and there are issues in practical use of the results of behavior recognition. In this study, we construct a system that can interpret actions by speech, and investigate a method to obtain a distributed representation of new microbehaviors. language of the presentation:Japanese | |||
| 澤野 耕平 | M, 1回目発表 | ユビキタスコンピューティングシステム | 安本 慶一, 渡辺 太郎, 諏訪 博彦, 松田 裕貴 |
|
title: Study of a location-based photo and text submission analysis method for discovering city objects of interest
abstract: Although city objects of interest, such as cherry blossoms in full bloom, are often absent from digital map services, this information is important for sightseeing navigation systems. In this study, we propose a method to discover such information by analyzing images and texts with location information posted on SNS and other services. language of the presentation: Japanese | |||
| 小竹 宙樹 | M, 1回目発表 | ヒューマンロボティクス | 和田 隆広, 松原 崇充, 織田 泰彰, 劉 海龍 |
|
title: Human-Robot Collaborative Carrying of Objects Using Robot Motion Information Presentation Device
abstract: Smooth task performance is required in collaborative tasks between humans and robots. In this study, we focus on the task of carrying fragile objects by a human and a robot. The load on fragile objects is limited and requires precise work. Generally, impedance control and other methods are used to limit the force applied to the object in such tasks, but understanding of the robot's behaviors is considered essential for humans to realize high-speed transport tasks from a human perspective. The aim of this research is to realize smooth transport of fragile objects by introducing a haptic device that feeds back the robot's state to the human. language of the presentation: Japanese 発表題目: ロボット運動情報提示デバイスを用いた人間・ロボット共同物体運搬作業 発表概要: 人間とロボットによる協働作業において円滑な作業実現が求められている．2エージェント間においてタスクを円滑に遂行させるためには，エージェント間の情報共有が必要である．本研究では人間とロボットによる脆弱物体の運搬作業に注目する．脆弱物体にかけられる荷重には制限があり，緻密な作業が求められる．通常，このような作業ではインピーダンス制御等によって物体にかかる力を制限する方策がとられるが，人間からみて高速な運搬作業を実現しようとする場合，人間はロボットの動作に対する理解が不可欠と考えられる．そこで本研究では，ロボットの状態を人間にフィードバックするハプティックデバイスを導入することで，スムーズな脆弱物体搬送を実現することを目的とする． | |||
| 酒井 大輔 | M, 1回目発表 | ヒューマンロボティクス | 和田 隆広, 松原 崇充, 織田 泰彰, 劉 海龍 |
|
title:Research on estimating machine maneuvarability and quantifying operator's proficiency using vestibulo-ocular reflex
abstract:Although research has been conducted on Human Machine Interface to improve the efficiency of machine operation, there are only a few studies to understand the mechanisms in the brain when humans operate machines. Assuming that "machines that are easy to maneuver can be controlled in accordance with predictions in the brain," we focused on the vestibular system, which plays a role in predicting movement in the brain. Motion prediction by the vestibular system is used in the vestibulo-ocular reflex (VOR) that stabilizes visual field. When operating a ride-on machine, the accuracy of VOR represents the difference between the prediction of the machine's motion in the operator's brain and the actual machine's motion. By investigating the relationship of VOR to the dynamics during machine operation, we aim to estimate the maneuvarability and quantify the proficiency level of operation. language of the presentation: Japanese 発表題目:前庭動眼反射による機械操作性の推定及び操作熟練度の定量化の研究 発表概要: 機械操縦の効率を改善するための Human Machine Interface に関する研究は行われているが，人間の機械操作時における脳内のメカニズムを理解する研究は数が少ない．本研究では，"操縦しやすい機械は操縦時に脳内の予測どおりに機械の動作を操れる"と考え，脳内における運動予測に役割を果たしている前庭系に着目した．前庭系による運動予測は前庭動眼反射（VOR）で用いられ，視界を安定させるために利用されている．搭乗型の機械の操縦時においてはVORの精度がオペレータの脳内における機械の運動の予測と実際の機械の運動の差を表すとして，機械操縦時のダイナミクスに対するVORの関係について調べることで，機械の操作性の推定や操作の熟練度を定量化を目指す． | |||
| LI SIYUAN | M, 1回目発表 | ヒューマンロボティクス | 和田 隆広, 清川 清, 織田 泰彰, 劉 海龍 |
|
title: Optimizing Visual Information for Readability and Motion Sickness in Autonomous Vehicles
abstract: The progress in automated driving has facilitated engagement in non-driving-related visual tasks, such as reading books and smartphones, leading to increased motion sickness and diminished readability. Numerous studies have explored approaches such as motion sickness models based on subjective vertical conflict (SVC) theory and providing visual information for motion sickness reduction. However, the effect of eye movement on motion sickness and readability was not considered in current methods. This study introduces a Vestibulo-Ocular Reflex (VOR) output path to the SVC model with visual input. Two methods are proposed for generating optimized visual information in response to anticipated VOR eye movements, serving purposes of enhancing readability and reducing motion sickness. The potential effectiveness of the proposed methods was demonstrated through simulations and will be further validated through ongoing experiments. language of the presentation: English | |||
| 今澤 拓也 | M, 1回目発表 | サイバネティクス・リアリティ工学 | 清川 清, 和田 隆広, 内山 英昭, Perusquia Hernandez Monica, 平尾 悠太朗 |
|
title: Virtual Visceral Floating Sensation
abstract: In roller coasters and other attractions where people fall from a high altitude, an upward force called "negative G" is exerted on them, causing their internal organs to float up and they perceive a unique sensation. An illusion similar to this sensation may be felt when viewing an image of falling from a height, which is referred to as a virtual visceral floating sensation. Vection is an illusion induced by visual information, in which a sense of movement of one's own body is induced despite the fact that the body is actually stationary. Although there have been many studies on vection and falling sensation, there have been few studies on virtual visceral floating sensation. This study aims to clarify more about the virtual visceral floating sensation. In this presentation, as a preliminary experiment, we propose an experiment to verify the effect of acceleration in video on the virtual visceral floating sensation and to compare it with vection. language of the presentation: Japanese 発表題目: 疑似的な内臓浮遊感覚 発表概要: ジェットコースターなどの高所から落下するアトラクションでは、マイナスGと呼ばれる上向きの力がかかり内臓が浮きあがり独特の感覚を知覚することがある。この感覚に似た錯覚を高所から落下する映像を視聴した際に感じることがあり、これを疑似的な内臓浮遊感覚と呼称する。視覚情報によって引き起こされる錯覚として、実際には静止しているにもかかわらず自己身体の移動感覚が引き起こされる錯覚であるベクションが挙げられる。ベクションや落下感覚に関する研究は多く行われているが、疑似的な内臓浮遊感覚に関する研究は少ない。本研究では疑似的な内臓浮遊感覚に関してより多くのことを明らかにしていくことを目指す。本発表では予備実験として、映像中の加速度の疑似的な内臓浮遊感覚への効果の検証およびベクションとの比較を行う実験を提案する。 | |||
| パニアグア カルロス | M, 1回目発表 | サイバネティクス・リアリティ工学 | 清川 清, 和田 隆広, 内山 英昭, Perusquia Hernandez Monica, 平尾 悠太朗 |
|
title: Tape-tics: Proposal for Tape Devices Arranged with Miniature Vibrators
abstract: Haptic devices using small vibrators are becoming more and more popular every year. In addition, research on vibration stimulation at a single point or over the entire body is also active. However, there are various problems with current haptic devices, such as limited presentation area and low scalability. We propose a haptic device that can be handled by non-experts regardless of the user. First, we introduce four studies that provide devices and tools that are highly scalable and support development. Next, we introduce the proposed method tape-tics. This is a device consisting of small LED tape-like oscillators connected in series. The user can cut the tape to suit his or her own preference and attach it to any desired location. Since this device is a tape device, it can be mass-produced at low cost. Then, progress will be described, and finally, future plans will be discussed language of the presentation: Japanese 発表題目: Tape-tics:小型振動子を並べたテープデバイスの提案 発表概要: 小型振動子を用いたハプティックデバイスは年々普及している.また、一点または全身に振動刺激を与える研究も盛んである. しかし、提示部位が限られていることや、拡張性の低さなど現状のハプティックデバイスには様々な問題がある．ユーザを問わず専門知識を持ったものでない人でも扱えるようなハプティックデバイスを提案する． まず初めに、拡張性が高く、開発を支援するデバイスやツールを提供している研究４つを紹介する。 次に提案手法tape-ticsを紹介する。これはLEDテープのような小型振動子が直列に接続されたデバイスである。ユーザーが自分の好みに合わせて切り取り、任意の場所に貼り付けることができる特徴がある。 このデバイスはテープデバイスであるため、安価に大量生産が可能です。 その後、進捗状況を説明し、最後に今後の予定について述べる | |||