| MA SHANPENG | M, 1回目発表 | インタラクティブメディア設計学 | 加藤 博一, 清川 清, 神原 誠之, 藤本 雄一郎, 澤邊 太志 |
|
Title: Improving Passenger Trust And Acceptance in Autonomous Vehicles through an Explainer Model
Abstract: This research focuses on enhancing the comfort and trust of passengers in autonomous vehicles by improving system transparency. Autonomous Driving Systems (ADS) are often perceived as a black box, making it difficult for passengers to predict the vehicle‘s behaviour. This uncertainty can lead to discomfort and reluctance to use autonomous vehicles. While most research on autonomous vehicles focuses on safety and efficiency, there is a lack of studies on passenger trust and acceptance. This study proposes an “Explainer” model that uses Augmented Reality (AR) to present structured data from the Autonomous Driving System (ADS) in a human-understandable format. The model categorizes data from the Perception, Planning, and Control modules of the ADS into causal and intentional information. Causal information presents the physical cause and effect relationship between the vehicle's movement and its environment, while intentional information presents the vehicle's target and action intent. By providing passengers with a clear understanding of the vehicle's behavior, this model aims to enhance their sense of safety and control, thereby improving their overall comfort and trust in autonomous vehicles. language of the presentation: English | |||
| ZHANG RENJIE | D, 中間発表 | インタラクティブメディア設計学 | 加藤 博一, 向川 康博, 神原 誠之, 藤本 雄一郎, 澤邊 太志 |
|
Title: Intelligent Display for Head Mounted Display
Abstract: In recent years, head-mounted displays (AR) have been increasingly recognized by many people as having the potential to replace smartphones and become the next generation of productivity tools. We analyzed the changes brought by past productivity tools and their underlying reasons, and based on the characteristics of head-mounted displays (AR), we proposed the concept of intelligent displays. In daily use, applications for HMD are expected to actively change the displayed content and its position based on the current environment, user behavior, and habits. However, existing research and applications on HMD mostly focus on improving the quality and effectiveness of displayed content within fixed environments. These applications often have limited functionality, lack interconnectivity, and struggle to provide an ideal user experience. The main reason for this problem lies in the difficulty of understanding the surrounding environment and analyzing user intentions, which making it challenging for these applications to adapt to complex and dynamic scenes. The goal of this study is to develop technologies that constantly analyze the surrounding environment and user intentions, thereby addressing the problem of "what to display, where, and when." Additionally, we are committed to providing datasets for the ongoing improvement of subsequent systems and contributing to the sustained progress in this field. Language of the presentation: English | |||
| ZHANG JIAXUAN | D, 中間発表 | ユビキタスコンピューティングシステム | 安本 慶一, 和田 隆広, 諏訪 博彦, 松田 裕貴 |
|
title: Construction of Human (Especially Child) Abnormal Gait Detection System Based on sEMG
abstract: Gait is a complex motor function characterized by the interaction of many body structures and the central nervous system. Gait analysis plays an important role in many medical and health fields. Since gait abnormalities begin child, early examination of gait abnormalities is important for predicting the risk of neurological damage, and preventing motor damage and arthritis. However, most of the current abnormal gait detection systems are oriented to adults and cannot be applied well to children. This research aims to build a gait abnormality detection system that is convenient and usable by children. We propose a novel 3-axis view of sEMG features composed of temporal, spatial, and channel-wise summary. We leverage deep learning technologies to enforce efficient parallel search and to get rid of limitations imposed by previous work in gait classification. Our model is designed on top of an attention-based module, which allows for the extraction of global contextual relevance among channels and the use of this relevance for sEMG recognition. We compared the proposed method against existing methods on two Ninapro datasets consisting of data from both healthy people and amputees and achieved new SOTA. language of the presentation: English | |||
| JAISRI PONGCHAI | M, 1回目発表 | ソフトウェア工学 | 松本 健一, 安本 慶一, 石尾 隆, Raula Gaikovina Kula, 嶋利 一真 |
|
title: How Library's Contributors Contribute Back to Each Other: A Study of GitHub Issues
abstract: Nowadays, there are many third-party software development. This led to large inter-dependency networks. These libraries must be maintained by their maintainers. For example, create pull requests, submit issues, etc. However, the maintainers of the large library can be overworked. So there is a group called contributors who assist the maintainers in keeping the libraries running. One of the contributors is the client's maintainers, and the library's maintainers also contributes back to the client. By analyzing GitHub issues, I investigate the relationship between the library and its clients in order to understand how the library's and the client's contributors contribute back to each other. Our findings can be used to better understand the main purpose of the issues submitted by library contributors to their clients, or vice versa. language of the presentation: English | |||