Recently, the integrated bigdata platform between social components is paid attention toward realization of the super smart city and society. In this paper, multiplex and integrated optimization of power system and transportation system was focused aiming at energy saving, peak load suppression, power regeneration and contribution to connected power distribution network system. The method to develop an integrated model of Light Rail Transit (LRT) and distribution system (DS) was proposed to analyze mutual effect on each system caused by operation of LRT. As a use case, Utsunomiya city was focused on as a city which aimed transportation electrification toward low-carbon city. The integrated model was constructed assumed the Utsunomiya LRT which was under construction and the DS of Utsunomiya city. Using this model, some effects were evaluated on the DS operation by LRT introduction and EV penetration. The peak load suppression effects on power demand of LRT were also simulated using the developed model by coordinately controlling EV chargers and LRT.
Societies in Japan face many social issues today. To address these issues, the government is exploring ways to solve problems using data. The purpose of this study is to provide support using existing data for grasping the current situation, taking up the problem of bicycle parking environment in a university as an example of social issues. Specifically, we tried to design a unique schema and build knowledge graph for integration of various on-campus data. Then we applied and modified the knowledge graph completion methods to improve efficiency and accuracy. For the knowledge graph constructed in this study, the number of bicycles was estimated by the conventional method and the proposed method, and the transition of the sum of absolute value errors was compared. The proposed method exceeded the existing method in efficiency and accuracy. In addition, approximately 650 data aggregated at any date and place were able to be estimated about 54 units correctly, compared to the conventional method. Finally, we improved efficiency and accuracy of knowledge graph completion methods.
The influence of foreign tourists on the Japanese economy is significant. However, the number of tourists have grown at a sluggish pace in local region. In local region, there are attractive features not found in urban areas, and there is possibility that hidden information can be transmitted in addition to the existing famous tourist information. Therefore, there is a demand from tourist operators to discover the potential needs of their area.
In this paper, we propose a method for analyzing objects that the area has given interest to visitors from images posted on the social networking service to discover potential needs in the region. The feature of this method is that it uses an image analysis service based on deep learning opened on the cloud service. This image analysis service can recognize the objects in a image and output their names as tag information. Therefore, by collecting SNS's images in any region, converting then into tag information, and statistically analyzing the tag information, the objects that the region is interested in is extracted as tag information. In this paper, we propose a analysis method of tag information based on the frequency of appearance tags and based on the difference between appearance tags in other regions. And the effectiveness was verified through the several experiments. In general, deep learning technology requires collection of large amount images and labeling for learning data in order to construct an image recognition model. In addition, in order to learn the model, it is necessary to prepare high spec computer environment. When introducing the analysis system of SNS's images, it is difficult to request them from tourist operators. On the other hand, the model of cloud service will be updated year by year, and the accuracy and versatility will be improved. Therefore, the effectiveness of using cloud services to analyze SNS's images is clarified in this paper.
While some experts have been talking about the decline of local communities, at present there are many charitable local activities for someone, for communities, and for society. However, some of inhabitants mainly elderly people living alone and child-rearing mothers cannot obtain information on their living area because they have few opportunity to interact with the local residents. In order to reduce person who are isolated from local community due to the lack of information and realize the new kind of relationship suitable for the current social system, the local community needs a mechanism where both the isolated person and the person contributing to community can easily recognize each other's existence and connect them. Therefore, we develop a smartphone application named “Tamemap” aiming at a society where all local people are informally connecting in the region. Tamemap provides a function by which users living the region can transmit and catch small activities called micro events in their area. From the result of the real operation so far, Tamemap had been able to support users mainly people of child-rearing generation and seniors who need local information. Also, through the information sharing of micro events, Tamemap had been able to support the activities of users connecting two or more independent local communities.
With the expansion of Society 5.0, the realization of a short-term development and secure implementation for a smart factory system is required. SROS2, an open-source platform applicable to industrial robots, realizes short-term development and implements secure communication using AES-GCM. On the other hand, AES-GCM is reported to be vulnerable in computational security. Therefore, a secure encryption alternative to AES-GCM and its evaluation are needed. For these reasons, this research incorporates SROS2 with a secure certified encryption alternative to AES-GCM. The evaluation experiments in this study performed on a computer and Raspberry Pi 3 Model B, and the results revealed the effectiveness of dedicated hardware implementation.
Data stream mining of IoT data can support operator to immediately isolate causes of equipment alarms. The challenge, however, is to keep their classifiers high purity (the data ratio with same proper class in a cluster) with concept drifting ascribed to differences between alarm models and entities. We propose to continuously update data class according to their distribution changes. Through evaluation, no purity deterioration was verified for oscillation condition data with a drifting rate of 1%. The result suggested that the method improves operator decision making.
Movement trajectories can provide useful information for all fields. However, they may have high privacy parameters, sharing trajectory data with other operators without anonymization carries the risk of linking movement trajectories to individuals. Therefore, it is necessary to consider applying privacy protection to trajectory data. Anonymization indicators, such as k-anonymity are generally adopted for anonymization of trajectory data. Some studies modify position information to achieve anonymity. This modification sometimes produces inaccuracies in data sets. Moreover, they cannot decide on which balance to change location information and timestamp. In this study, to reduce the modification distance of the position, we propose an algorithm that allows the mismatch of time when the position information is acquired within a certain range. Further, we define indicators that represent distortions of position and time information. As a result of comparing the proposed method and the existing method, the usefulness of the proposed method is shown.
Newspaper in Education (NIE) is an educational approach to utilize the newspaper as a teaching material. The purpose of our research is to support elementary school teachers by recommending Web news suitable for regional study in NIE for elementary schools. This paper analyzes news articles within NIE worksheets and proposes a method to determine news articles suitable as teaching materials for regional study using Support Vector Machine (SVM) based on features of the contents of news articles, regionality and readability of articles. The effectiveness of the proposed method was evaluated by the experiments. From the experimental results, we confirmed that the proposed method using Word2Vec, Learned Kanji and Local Characteristic Words as the features obtained the highest precision of 87.6%. We also confirmed that there is a possibility that the proposed method can be applied to news articles in any region.
Collaborative design is a suitable approach to obtain useful and practical results within a group of people in a relatively short period of time. Our focused problem is to build a large-scale poster for promoting our college and publicizing entrance examination; its mission is important, therefore, we need good performance and accountability in the process of collaboration for product design supported by Kansei Engineering, because of its capability. Since the design process can involve complex decision-making variables, traceability is important to validate the consistency of decision making toward results design. We decided to employ Kansei Engineering to achieve a respectable performance of collaborative design and to apply it to Web system for decision-making. This paper describes our approach in performing collaborative design through our Web system and in demonstrating confirmation of consistency in evaluating product design based on collaboration using the Analytic Hierarchy Process (AHP) methodology.
The aim of this study was to assess Virtual Reality (VR) based visual auditory stimulation can relieve pain. To achieve this, the impact that body surface electrical stimulation has on the body surface potentials and activity in the central nervous system in VR environments were extracted. Twenty-two adults participated in this study. The body surface potentials and Electroencephalogram frequencies were measured under various conditions: non-viewing, viewing neutral clips, viewing unpleasant clips, and viewing pleasant clips. The pain index was calculated from the body surface potentials. Participants were asked to complete subjective pain measurement scales. Results showed that the pain index and the relative subjective pain score decreased when presented with unpleasant and pleasant video clips. The incidence of alpha waves declined when presented with unpleasant clips, while the incidence of beta waves increased when presented with unpleasant and pleasant video clips. Inattentiveness to pain due to VR immersion and descending pain inhibition due to awareness of emotional stress after viewing VR clips were considered to have contributed to the decline in the pain index and in relative subjective pain score. This study suggests that it is possible to relieve pain by VR viewing related to unpleasant or pleasant.
PV (photovoltaic) generation has experienced the most growth among the renewable energy sources in the last several years, and the interconnection capacity of PV generations may continue to increase. However, more recently special problem in Japan, the capacity of qualified PV generation facilities exceeds the allowable interconnection capacity determined by each power grid company. This fact is recognized as a critical issue for reliable operations of power grids, and it will be enforced by law that the power grid company can announce the output power curtailment instruction signal to the PV generation plants. An output power curtailment essentially means wasting potential solar energy. Operating methodologies for the PV generation plant that can reduce the amount of wasted energy are needed to be developed. This paper considers the PV generation plant equipped with the energy storage and investigates decentralized management of power conditioning systems which are used to interconnect the PV and storage system into the grid. The effectiveness of the proposed methodology is evaluated through the real physical experiments.
Corresponding point searching is the most important process at the passive stereo measurement. At the long-range target, the number of pixels of binarized target marker images are sometimes few, concretely about 100 pixels. In this case, the fluctuation of the coordinate of the corresponding points is possibly appeared. We have been studying long-range outdoor measurements by the high-precision passive stereo to detect the small movement of the land for the disaster prevention. For this application, the stability of measurement results is required to avoid the mistake alert.
In this paper, we discuss the fluctuation in detail using several shapes, sizes and rotation conditions. Also, we found the condition that showed outliers which must be avoided for the image measurement. As a result, the circle, the most popular marker shape, sometimes shows more outliers than some of the other shapes such as the slightly rotated square marker.
Nowadays, natural images in the real world can be affected by more than one type of noise during the process of image acquisition and transmission. Many researchers have been trying to remove the mixed noise because it is a major problem to be considered in image processing. Therefore, they utilized many denoising methods for removal of mixed noise because different noises have different properties. In this paper, we consider the removal of mixed noise composed of Additive White Gaussian Noise (AWGN) and Random-Valued Impulse Noise (RVIN). Although most mixed-noise removal methods can successfully suppress the noise, some image details are lost in edge and texture regions because of the miss-detection of the image details as the impulse noise. Hence, we propose a mixed-noise removal method to preserve the image details. Our method is divided into two steps. The first step is to estimate the denoised image by integrating interpolation, DWM filter, down-sampling and BM3D. The second step is to preserve the image details lost in the first step by calculating the absolute difference between the input noisy image and the pre-estimated image obtained from the first step. The core of this paper is that the input noisy image is initially interpolated by multi-surface fitting for single frame before impulse noise detection of DWM filter in the first step to maintain the image details in the edge and texture regions. Experimental results show that our mixed noise removal method is superior to the state-of-the-art image denoising methods in terms of both quantitative measure and visual perception quality.
As a parameters optimization method for neural networks which is applied to reinforcement learning, Evolution Strategy has been proposed. In this method, neural network parameters are represented by individuals, like ordinary evolutional strategies. While the evolution, a new individual is generated from some distribution that centered a parameter and is weighted according to the order of reward that the neural network corresponding to the individual obtained. However, there are cased that the differences of reward values among the higher order individuals are so few that the updating can not lead to individuals to higher quality. So, in this research, after updating the normal parameters, we select the top individuals who get high rewards and weight them, and propose a method to update the parameters again using those individuals. By focusing on individuals who get a high reward, it is expected to search for a parameter that can obtain a high score earlier than the conventional method. In the experiment, the conventional method and the proposed method are applied to BipedalWalker which is a learning environment of a 2D biped robot in OpenAI Gym, and evaluation is performed and as a result, the proposed method showed better performance than the conventional method.
Lettuce leaves grown in a plant factory were photographed by a digital camera and a near-infrared camera. We propose an image processing method to investigate the differences in growth depending on the kinds of lettuces and the kinds of LED irradiation. The normalized difference vegetation indexes (NDVI) for the leaves were also measured. We show the effectiveness of the proposed image processing method as nondestructive measurement.