It is necessary to support of computer operation for a physically disabled person. One of the possible physical movements of the physically disabled person is Eye movement. Eye gaze and detection of eye opening and closing of a person makes it possible to operate a computer. In our system, a web camera, Gaze pointer, Dlib C++ library and OpenCV library are used to recognize eye gaze and detect eye opening and closing. Changing the eye gaze, we can move a mouse cursor. Recognizing closed eye, we can carry out an operation of mouse click. In this paper, we evaluated the effect on operability due to the eye gaze and detection of eye opening and closing.

抄録全体を表示

This paper deals with the theme of the development of story units based on Japanese folktales. Each story unit functions in an Integrated Narrative Generation System (INGS), which is the core of our systematic narrative generation study. In this research framework, story units can be positioned as a story technique that generates and expands a story structure in the narrative generation process. This paper focuses on the combination of story units implemented by Common Lisp using a verb conceptual dictionary, which is important for a noun conceptual dictionary in conceptual dictionaries functioning in the INGS. The mechanism enables organically combined processing with other narrative techniques and the use of semantic functions using conceptual dictionaries functioning in the INGS.

抄録全体を表示

The goal of software testing is to detect all latent defects. However, it is difficult to know how many latent defects remain and where they are hidden. In this research, we propose a system that analyzes the characteristics and tendencies of already detected defects and predicts where the latent defects are likely to be. Specifically, the system inputs the data of detected defects into a Self-Organizing Map (SOM) and predicts the locations that contain many defects from this map. To confirm the validity of this proposal, we input past defect data into the SOM, analyze the trend of defects, and evaluate the predictability of the latent defects.

抄録全体を表示

Our previous study developed a testing and training device, “iWakka,” for the adjustability for grasping force. Moreover, we developed “iWakka Game” for autistic patients. In this study, we investigated the applicability of the iWakka Game in children with developmental coordination disorder (DCD). We developed a small grasping body and improved its evaluation method. It was experimented on four children with DCD. Consequently, they completed the evaluation task, which enabled the extraction of hand dexterity characteristics.

抄録全体を表示

Assistive devices have been studied in the senses of the universal design and a possible simplification for mass production, and such a concept can be extended to the personalization by using advanced 3D printing technologies, which allows to provide suitable exoskeleton-type devices to fit for the subject’s body to achieve symptom relief. A systematic analytical mechanics is highly important for the investigation of proposed integrative systems and multibody dynamics and its computer simulation enhance the capability to find potential problems in design, for example risky postures for the subject. We introduced multibody dynamics to analyze a knee support exoskeleton by using a linkage system newly designed as a simplification of the device. In the computer experiment, a singular posture was successfully found as the limitation of the device to move safely for subjects having joint dysfunctions.

抄録全体を表示

This paper proposes Gamma, which is a messaging platform for spatial data. Gamma uses distributed MQTT system for spatial data sharing, and this paper aims to improve the usefulness of distributed MQTT system. In the experiment, it has been found that Gamma can process more messages than a single MQTT broker. It is confirmed that Gamma achieves scalability by increasing the number of Gateways and distributed MQTT brokers. Furthermore, it is confirmed that the efficiency of the distributed MQTT system can be improved by setting the responsibility area of Gateways.

抄録全体を表示

Improving the efficiency of fish feeding contributes to achieving sustainable expansion of the aquaculture industry. However, expert knowledge on feeding remains reliant on experience. This paper presents a new approach of applying digital transformation (DX) on such knowledge by measuring underwater currents induced by fishes as indicator of their hunger. A prototype sensor suite consisting of current sensors, cameras, and an inertial measurement unit (IMU) was constructed to measure the currents around the fish cage, measure the motion of the suite, and record fish activity, particularly during feeding. An initial experiment was performed in two fish cages during feeding activity. Current measurements were collected and analyzed together with the recorded videos to relate the changes in current to the feeding activities. Results suggest that the currents increase around when feeding was started continuously and decline to zero around when it was stopped. However, more data needs to be collected and analyzed for a better understanding of the relationship between the changes in current and fish activities so it can be used to optimize the feeding decisions of fish farmers.

抄録全体を表示

In this paper, we propose a work revision method using reinforcement learning on a constructive nurse scheduling system. The constructive nurse scheduling system makes the shift schedule creation procedures and rules easy to understand, because the system does not use the evaluation value for the entire shift schedule. We have confirmed the possibility of improving the quality of the shift schedule by the proposed method through a computational example.

抄録全体を表示

Emotion recognition is useful in many applications. Most methods available nowadays depend on facial features which are difficult to obtain from standard security cameras. Unlike traditional biometrics, gaits can be obtained noninvasively from afar. A novel method to perform gait data collection with real-time emotion induction was proposed. OptiTrack was used to capture 49 participants walking circularly while watching emotion-induced videos on Microsoft HoloLens 2.This is the first study that use real-time emotion induction technique in non-straight walkingpath.

抄録全体を表示

This paper proposes ternary or binary weights with 8-bit integer activation convolutional neural networks. Our proposed model serves as the middle ground between 8-bit integer and lower than 8-bit precision quantized models. Our empirical experiments established that the conventional 1-bit or 2-bit only-weight quantization methods (i.e., BinaryConnect and ternary weights network) can be used jointly with the 8-bit integer activation quantization. We evaluate our model with the VGG16-like model to operate with the CIFAR10 and CIFAR100 datasets. Our models showcompetitive results to the general 32-bit floating point model.

抄録全体を表示

Less and less Japanese are working in the forestry sector. Due to this lack of workforce, dangerous tasks could be delegated to robots. In order to perform this mission, the latter need embedded localization and mapping systems. This paper discusses the implementation of a particle filter based SLAM solver on a dedicated mobile robot called SOMA, developed at Hayashi Laboratory. The proposed approach has been evaluated in a realistic forest simulation and the first results suggest that it could be used in real time

抄録全体を表示

The delta robot synchronous control based on the Ethernet Control Automation Technology (EtherCAT) protocol for painting application is proposed in this paper. Personal Computer (PC) is used as master and the delta robot motor drivers are used as slaves in this work. The Master sends command to slave base on the motion control profile CAN in Automation 402(CiA402). Subsequently, the program in C# perform the user’s interface and EtherCAT communication. And the system is not only for easy use, but also for quickly high-precision. A complex painting application is proposed to show this system workable.

抄録全体を表示

The main objective of this study is to probe how Taiwanese building investment and development companies rate the analytical framework and weights of artificial intelligence buildings. Document Analysis, the Delphi method, and the Fuzzy Analytic Hierarchy Process (FAHP) are applied to conduct a FAHP questionnaire survey among 20 building investment and development companies in Tainan. Based on the calculation of composite weights, the findings are: (1) The most crucial evaluation indicator for security and hazard prevention is the “access control system”. (2) The most crucial evaluation indicator for energy-saving management is “energy-saving technology”. (3) The most crucial evaluation indicator for health and comfort is the “interior comfort system”. (4) The most crucial evaluation indicator for intelligent innovation is the “intelligent innovation concept”.

抄録全体を表示

Paintings can be damaged by natural causes or accidents. One of the crucial natural damages was frequently caused by mold defects. The mold discovery is an important step in the restoration of damaged paintings. The procedure is usually tedious and depends heavily on the qualitative visual judgement of an expert restorer. The aim of this work is to assist the restoration process via an automatic mold defect detection technique based on derivative and image analysis. This new method, designated as Derivative Level Thresholding (DLT), combines binarization and detection algorithms to detect mold rapidly and accurately from scanned high-resolution images of a painting. This work also benchmarks the performance of the proposed method to existing binarization techniques of Otsu’s Thresholding Method, Minimum Error Thresholding (MET) and Contrast Adjusted Thresholding Method. Experimental results from the analysis of 20 samples from high-resolution scans of 2 mold-stained painting have shown that the DLT method is the most robust with the highest sensitivity rate of 84.73% and 68.40% accuracy.

抄録全体を表示

Internet of Thing (IoT) and robot technology is a very popular research topic. Especially in the application of Industry 4.0, it is the most basic and important part. However, related security and application system development often have many problems. For example, the integration of various communication interfaces and data formats, and the rapid establishment of application systems. Therefore, this paper uses MQTT protocol and AES encryption technology to develop a Web-based SCADA system. This Web-based SCADA uses drag-and-drop operation, and users can quickly build a WYSIWYG (What You See Is What You Get) application system. Moreover, with the characteristics of MQTT message transmission, users can choose to transmit the device's message as it is or to process the data before transmitting when creating the SCADA system. And these messages will be encrypted by AES, making the whole system safer and more efficient. Finally, by collecting different communication protocols and integrating multiple communication interfaces, this SCADA can be connected to many PLCs and other equipment, so that the industry can build application systems quickly and at low cost.

抄録全体を表示