Many people are indecisive and take a long time to decide on a dish to eat at a restaurant. Therefore, in this study, we focused on deciding which dishes to eat and developed a system that suggests dishes to eat by entering the parameters of meal preferences (UP) in advance. In this system, the developer sets the parameters (DP) of the ingredients used and the main eating time for each dish. The system then compares the DP based on the UP and the current time (TP), and gives each dish a score, and suggests to the user the dish with the highest value．In this system, not only individual but also group suggestions were made.

View full abstract

In this study, we propose a new method for synthesizing non-photorealistic images by applying style transfer technology. Our method involves extracting features from different style images and combining them to apply to a single content image, thereby creating images with unique artistic effects. The experimental results demonstrate that our method effectively merges multiple styles and produces images with high artistic value. This research opens new possibilities in the fields of image processing, digital art, and computer vision.

View full abstract

In this paper, we construct a simple deep learning model for quickly transporting objects that arrive at irregular times, and for adjusting them at a certain period (or it's positive integer multiple) by using n conveyor belts arranged in a row. Previously, we proposed a method for determining the speed of each conveyor belt by iterative calculation of Linear Programming, but there was a trade-off between optimality and the amount of calculation. Here, we treat the Linear Programming based algorithm as an input-output system and show that it can be replaced by the simple deep learning model with sufficient accuracy. As a result, we show that the speed command value for one object can be generated in less than about 1/1000 of the time required by the conventional method, and that throughput is improved.

View full abstract

In this study, by using a cover without thickness, as maximum capacity, a tip-resistant trash box even if load forces act at the end was proposed. The thickness that maximizes capacity of the formed trash box was found, and the maximum capacity of the formed trash box was found from the found thickness. The maximum capacity was large in case that width is larger than height in the cover, a flap was on left and right in long and rectangular formats, and a flap was on up and bottom in western formats. The trash box became harder to fall down, because fall load forces in the length and thickness directions were large in case that the width is larger than the height in the cover. Therefore, by using the cover that the width is larger than the height, as the maximum capacity, the cover was the tip-resistant trash box even if the large load forces act.

View full abstract

In recent years, rapid changes in the global environment are also affecting marine ecosystems. For example, ocean acidification and rising sea temperatures are said to make echolocation difficult for marine mammals. A particular problem is that sea urchins eat all the seaweed. Seaweed plays an important role for marine life. Therefore, in this study, we developed a device to efficiently collect sea urchins. Devices for collecting sea urchins had previously been developed. However, there were issues with their suction power and size. In this study, we developed an underwater device that applies the ejector principle and maintains the same suction force. This device does not require a pump, which allows underwater robots to be made smaller and lighter. In addition, it is thought that this system could also be applied to collecting dead fish at fish farms. Currently, divers periodically dive underwater to capture the dead fish. Therefore, we also conducted experiments to confirm the effectiveness of this system in collecting dead fish.

View full abstract

This study analyzes tourism-related data in Nagasaki City and proposes a method to examine how various information sources, such as newspapers, television, websites, and social networking services (SNS), influence travelers to information about restaurants, events, experiential programs, facilities, and transportation. Additionally, a method for evaluating the degree of adequacy or insufficiency of information for travelers is proposed. The analysis was conducted using the results of an online questionnaire survey conducted by Nagasaki City targeting tourists. The findings reveal that information sources with a minor influence on travelers during their trips include "transportation company websites (railways, buses, airlines) ", "personal blogs", "video-sharing platforms (e.g., YouTube) ", and "newspapers". Conversely, information sources with a significant influence include "travel guidebooks", "family, relatives, and acquaintances", and "pamphlets obtained during the trip". Furthermore, the study identified that information about "restaurants and gourmet options", "travel routes and transportation", and "costs" is insufficient during travel, highlighting the need for specific improvements in these areas.

View full abstract

We develop a system (attention function evaluation system) to evaluate attention function by reaching targets using image processing. The attention function evaluation system does not use devices such as controllers, but uses an interface (spatial switch) that draws the targets on an image captured by a camera and judges contact when the subject blocks the target. To verify the effectiveness of the attention function evaluation system, we will conduct a comparative experiment with conventional testing methods on elderly people with MCI (Mild Cognitive Impairment) and with mild to moderate general cognitive decline. As a result of the experiment, by using the attention function assessment system, we were able to measure the direction of attention and the time required for the reaching movement due to spatial switching, which could not be obtained using the conventional method, and we were also able to confirm the relationship between attention function and cognitive function.

View full abstract

No functional testing methods to observe the recovery process of patients undergoing treatment with an ileus tube have yet been proposed. In this study, we developed a system that measures small intestine motility as changes in balloon pressure attached to an ileus tube during treatment, using a pressure sensor. As a preliminary technical evaluation of the system, we conducted measurements on five healthy volunteers: three received metoclopramide to promote motility, and two received butylscopolamine bromide to suppress it. We assessed whether the system, along with its signal analysis methods—specifically, time-domain and time–frequency analysis using wavelet transform—could quantitatively detect drug induced changes in motility. After metoclopramide administration, pressure fluctuations increased by approximately 0.2–0.4 kPa in the 1–2 cycles/min range, along with increased spectral power and activity. Conversely, approximately 27 seconds after butylscopolamine bromide administration, a pressure drop of 0.5–1.0 kPa and decreased activity in the 1–2 and 6–11 cycles/min ranges were observed. These results suggest that the system, combined with appropriate signal analysis methods, is technically capable of quantifying motility changes and may serve as a useful tool for clinical monitoring.

View full abstract

This paper describes a system for detecting anomalies such as outliers and periodic fluctuations in time-series data, including data acquired by satellites, and how to improve the accuracy of the system. Although there are various methods for anomaly detection, machine learning methods have been actively studied in recent years, unsupervised learning methods that do not provide correct labels are rare due to their low accuracy. We developed an anomaly detection system based on unsupervised learning using TadGAN, an adversarial generative network, and solved its drawbacks by performing hyperparameter tuning using FlexHyperband. This system enables fast anomaly detection on large data sets without the need for human intervention.

View full abstract

One of the measures to achieve the 2050 carbon-neutral goal is to promote the use of household cogeneration systems powered by hydrogen fuel cells. However, this has not progressed as planned due to the high cost of fuel cells. As a countermeasure, a hydrogen engine that can replace fuel cells is a low-cost alternative, but its practical application faces the problems of reduced thermal efficiency and abnormal combustion, which are concerns due to the properties of hydrogen. In this report, the thermal efficiency was investigated using an engine suitable for household cogeneration, and extremely high thermal efficiency was obtained.

View full abstract

In this study, to promote resident-oriented community development, we analyze the generational differences in community awareness and clarify that community awareness differs between generations. Specifically, we conduct a questionnaire survey regarding community awareness in Otaru City, and conduct a goodness-of-fit test based on the questionnaire survey to analyze differences in community awareness between generations. At this time, the questionnaire survey included 48 questions and was analyzed in four age groups: "teens", "20s/30s", "40s/50s" and "60s/70s and over". As a result of the analysis, it was found that the younger generation has a higher awareness of the community than older people.

View full abstract

The authors have proposed a child monitoring system capable of estimating four risk conditions during a child’s outing-lingering in dangerous locations, deviating from designated school zones, abduction by vehicle, and immobility in arbitrary places—and notifying the guardian’s device of such conditions. The novelty of the proposed system lies in its foundation on the Crime Opportunity Theory and its ability to quantify the degree of danger at any given location. This enables more flexible condition assessment compared to conventional system. In an experiment conducted using a prototype within an actual school district, the simulated child device successfully inferred all four risk conditions and transmitted the corresponding notifications to the simulated guardian device.

View full abstract