Assistive technology (AT) is any item, device, software, or product system used to enhance, maintain, or improve the functional abilities of people with disabilities. To support these disabled and elderly people, we formed the National KOSEN Support Equipment Development Network (KOSEN-AT) with faculty members of technical colleges 10 years ago. In parallel with the above background, In Japan, the National Institute of Technology (KOSEN), started its new research project called GEAR 5.0 in May 2020. The AT and medical engineering are one of those key disciplines selected as GEAR 5.0 project. In this session, the special issue introduces the mission of KOSEN-AT in the GEAR 5.0 project, the newly proposed concept of Extended AT, and the research seeds of GEAR-AT to foster advanced future engineers.

抄録全体を表示

Wheelchairs are welfare support devices that are indispensable for the lives of people with disabilities, including the elderly who have lower limbs problems, and contribute significantly to the QOL of users. Manual wheelchairs can move in the front-back direction and turn, but cannot move in lateral or diagonal directions. Therefore, it is difficult to operate them in a narrow space. On the other hand, many studies on omnidirectional movement mechanisms have been studied so far, but in order to utilize these functions, it is essential to transfer from a manual wheelchair a person is using. In order to solve this problem, we have developed a removable omnidirectional automotive mechanism that can be attached to the wheelchair and detached from it instead of changing wheelchairs. In this paper, we show the configuration of the proposed omnidirectional movement module and the experimental results of the omnidirectional movement accuracy of the first prototype. In addition, we outline the second prototype, which is an improved version of the first prototype for social implementations.

抄録全体を表示

With the development of information and communication technology, practical dialogue systems that can talk with humans are attracting a great deal of attention. Dialogue systems are broadly divided into two types: task-oriented and non-task-oriented. The latter is aimed at dialogue itself such as chat, and basically it is necessary for the user to provide a topic, there is a problem that the motivation for dialogue is reduced due to boredom and familiarity. Therefore, a non-task-oriented attentive listening dialogue system is proposed to improve the problem. In the system, a new method is adopted that combines a response generation model with BERT as an encoder and Transformer as a decoder and transfer learning. In-depth questions and empathy responses are generated to realize listening dialogue by learning the dialogue data acquired using the Twitter API. The process of increasing vocabulary was reduced and response times were shortened and natural dialogue became possible by learning the huge amount of dialogue data collected from Twitter. In addition, by using transfer learning, it became possible to control the response style, also the problem of reduced dialogue motivation caused by poor topic was improved.

抄録全体を表示

This study is intended to produce a learning support system for middle-aged and elderly people with acquired visual disabilities. Our investigation results related to learning tools show that the use of Braille and PCs has decreased in the class of special support education. The primary reason is that finger reading is difficult for people with acquired visual disabilities. This paper proposes a function to read the text in the position by finger tracing on any screen from a textual document that is displayed on the tablet screen. The purpose of this study is to develop a smart terminal that reads the text information in the finger tracing location without using a digital recording machine or the like. This function is implemented in smart terminals to support the reading and writing of the visually impaired by the medical education institution and aims at practical use as a writing support device. This study is intended to produce a learning support system for middle-aged and elderly people with acquired visual disabilities.

抄録全体を表示

This study deals with smartphone application to support the walking of visually impaired. It recognizes surrounding situation from image acquired by camera and notifies it of any danger. The visually impaired has also a high desire to go out, however there are so many situations outdoors where they feel danger. This study used CNN (Convolutional Neural Network) which is a kind of deep learning techniques to detect obstacles and implemented into an application program working on smartphone. It can classify the taken images into some obstacles or surrounding situations and notice it to user to help his/her safety walking. This paper describes about the smartphone application and shows its effectiveness with experimental results.

抄録全体を表示

In this paper, it is reported on the current status of our efforts since last year, including the outline of the platform for synchronous collection of data output from four cameras for capturing the surrounding environment of the vehicle and the driving behavior of the driver, and one IMU sensor installed in the vehicle, as well as the results of data collection and analysis assuming a pedestrian running out as a verification experiment on campus. Based on the needs of medical professionals who have extensive experience in making medical decisions regarding the driving of senior, the objective of the development was to simplify measurement and analysis through sensor fusion. The developed platform uses four Raspberry Pis and a communication system called ROS (Robot Operating System) 2 to link them together. In addition, the platform includes a user interface system to enhance the convenience of operation using the M5Stack Core2. The experimental results show that our platform is inexpensive, highly customizable and synchronized.

抄録全体を表示