An unpredictable random number generator (URNG) adopts a deterministic algorithm with volatile internal states of a microprocessor, which makes the output of the URNG practically unpredictable. This study examines the URNG design proposed by Suciu et al., wherein performance counters are considered as entropy sources. Our experiments confirm that the URNG with performance counters requires a relatively long sampling interval with a background task to produce a high-quality random sequence. On this basis, we propose a new URNG design that is suitable for embedded systems. A simple 128-bit LFSR (Linear Feedback Shift Register) is built in a processor, whose lower 32-bit value is used as a random number. If an adequate sampling interval is maintained, the derived values pass the DIEHARD test.

In our previous paper, the authors proposed a fault tolerant system that adopts field programmable gate arrays (FPGA) with dynamic partial reconfiguration (DPR), based on autonomous control of reconfiguration. This study presents an experimental implementation of the proposed system that utilizes the DPR feature of Xilinx Zynq-7000 SoC. The control logic of DPR is implemented as a Linux software on the embedded ARM processor of Zynq-7000. DPR is invoked via PCAP, which is the dedicated interface for the embedded ARM processor. Four tiles (reconfigurable areas) are prepared and dynamically reconfigured to avoid the firm error of SRAM-type FPGAs. An experimental fault-tolerant system with triple redundancy and logic roving is implemented, and the measurement results of the reconfiguration time and data transfer time are presented.

In this study, we developed a small prototype of a one-degree-of-freedom portable active suspension for welfare support devices. The suspension is expected to be applied to welfare support devices such as wheelchairs, transport trolleys, and child seats. This paper presents the mechanism and the state space representation of the developed suspension, and an evaluation of the frequency domain characteristics and time domain response of the controlled system with poles assignment method. Furthermore, experiments under three road surface conditions were conducted to verify the effectiveness of the developed suspension.

Achieving practical and full-scale use of drones will require a transition from a first-person view (FPV) flight based on visual radio control to an autonomous wide-area, long-distance flight. However, technology that enables drones to fly autonomously over a wide area and long distances while feeding back positioning in ever-changing real-world environments is yet to be established. We aimed to develop a SLAM system that combines ORB-SLAM and dense point clouds in the environment. The result of an evaluation experiment of the developed SLAM system using a simulator indicated that the estimated self-position was corrected using matching with the dense point cloud, and a system combining ORB-SLAM and the dense point cloud was developed. It was confirmed that a SLAM system combining ORB-SLAM and some dense points in the environment could be developed. We have achieved good results in basic operation trials, demonstrating the potential of both systems for practical use.

This paper presents a novel seeds-needs cooperation method using an ICF model and a simplified embodiment model to develop nursing care robots. In the V-process of the nursing care robot development guidebook published by AMED, an ICF model is used to consider role of robots in providing nursing care. In addition, the simplified embodiment model contributes to the realization of robots to achieve better nursing care using the proposed method. In particular, the ICF model and the simplified embodiment model are useful to design and refine the development concept of a nursing care robot. The validity of the proposed method is confirmed by a development record of an all-in-one nursing care robot for transferring and moving support in the Okinawa nursing care robot consultation.

In 2019, we proposed an industry-school education skill succession (I-SSS) process and used it to develop a fisheries and marine skill learning system. In this study, we evaluate the appeal of the teaching materials in this system by using the ARCS motivation model from the perspective of learning motivation and search for the components of this system that would motivate learning. We used a teaching material evaluation sheet based on the ARCS motivation model to conduct a survey of students at a fisheries high school after using this system. From the results, we obtained many positive answers with all ARCS model factors, which confirmed that our proposed system had a high degree of usefulness for students. In addition, the results of the factor analysis on the motivation factor for student's learning show that the following the components lead to increased motivation for learning: “Procedure confirmation image, Presentation of finished form”, “Basic skills listed in textbooks”, “Explanation of practicality of skills”, “Explanation of the relevance of skills and work/Qualification”, and “Adoption of terms in textbooks, Easy-to-understand expressions”.

In this study, we aim to develop a robust motion recognition system for an intelligent video surveillance system, that can be used for security, sports and rehabilitation by using extended alternative learning. A robust motion recognition system is necessary for the automated detection of security incidents by using a machine learning approach. However, to avoid the difficulty of collecting a huge training dataset, we propose an alternative learning approach that trains a deep neural network with a 3D-CG dataset to recognize several motions. We present our experimental results on motion recognition from free-viewpoint videos by using deep learning and alternative learning. The trained deep neural network (DNN) is evaluated using actual videos by classifying the different actions performed by real humans in these videos.

This paper presents the development of a new method for the estimation and resolution of body occlusion using deep learning for an advanced intelligent video surveillance system. A generative adversarial network is used to estimate and reconstruct an image of a hidden part of the human body. Furthermore, an alternative learning approach using 3DCG that was developed in our previous study is adopted to create a large dataset for deep learning. Experimental results indicate that the proposed method performs well in the estimation of hidden parts of the human body using images of actual people.

A method of improving accuracy of UKF-based self-localization through the proper evaluation of error covariance matrices and its effect over improvement of self-localization of autonomous vehicles are suggested. First, error covariance matrices of spatial coordinate transformations defined as 1) an inverse of a coordinate transformation whose error covariance matrix is known, and 2) a synthesis of two coordinate transformations whose error covariance matrices are known, are derived. Second, a vehicle with a camera attached to a movable part favorable to landmark detection is presented. We demonstrate a case in which the accuracy of landmark-based self-localization of a vehicle is improved by using a camera fixed to a movable part that allows better tracking of landmarks, such as a steering rod of a micro-electric vehicle, if the error covariance matrices to be fed to UKF reflect the mechanical noise properly.

In recent years, damage by wild animals has become a serious problem, primarily in rural areas in Japan. In 2018, the amount of crop damage was estimated at approximately 15.7 billion yen. As a solution to this problem, a wild animal detection and expulsion system by drone has been devised, however, there are problems where the FPV operation requires human cost and the drone is operational only in a narrow area. It is essential that autonomous wide-area long-distance flight is required to realize the practical and full-scale use of drones. The authors have been developing a non-GPS autonomous control drone system for power line inspection, animal detection, and tracking. In this study, we propose a detection and tracking system with autonomous wide-area flight drones to prevent damage by wild animals. We verified the effectiveness of our combination of a deep learning system trained by images taken by a drone equipped with a camera having a wide field of view and flying at low altitude for image magnification, and our centerline extraction method for more accurate discrimination of wild boars.

We developed a fishery and marine skill learning support system as a self-training teaching material. In this study, we evaluated the usefulness of the proposed system by using text mining and searched for the points to be improved in this system. Using this system, we conducted a survey among the students of a fisheries high school using an evaluation sheet with free-form answering. From the results, we extracted five points, namely “learning method using this system”, “video operation and technology confirmation”, “promotion of understanding by this system”, “skill learning at home”, and “fun to learn fisheries skills”. Among these five points, “promotion of understanding by this system”, in which an understanding of the skill is promoted through explanations of skill and skill videos, was evaluated positively. It was also found that a review is effective in utilizing this system. As a point to be improved, it was found that narration should be provided along with the presentation of the subtitles.

The grade of a mango is determined by its color. Traditionally, the classification of mangoes is manually performed by humans. Such classification is subjective and then the quality of mango's is not uniform. In order to overcome the problem, automatic classification method to classify mango based on RGB histogram of their images and certain threshold. However, this process might be influenced by variations in the luminance. In this study, we implemented alternatives of the traditional method and compared them with sampled images.

To perform finite element analysis (FEA) for estimating the characteristics of synchronous machines (SMs), a current condition corresponding to an operating point is necessary as an analysis input. Accurate identification for the current condition is a difficult problem because it is strongly susceptible to magnetic saturation, e.g., the magnetic interaction between d- and q-axes. Especially in a wound-field SM, its flexibility makes the identification more complicated; in other words, three independent variables, such as the amplitude and phase of armature current and amplitude of field current, have to be identified for SMs, whereas two variables are required for a permanent magnet SM or synchronous reluctance machine. Thus, numerous researchers have studied construction of the identification method. This paper proposes a novel identification method for the current condition of wound-field SMs using saturation functions defined by the flux maps. It is demonstrated that the proposed method can correctly identify a large amount of current conditions in a short computation time. In addition, the influences of magnetic saturation on the characteristics of the SM were investigated by comparing the identification results obtained by the linear and non-linear FEA. It is revealed that the magnetic saturation drastically affects the current condition and must be considered in the identification methodology.

This paper proposes a capacitor-less AC-AC converter for a variable-speed wind power generation system (VSWPGS) using a switched reluctance generator (SRG). As a new type of converter, the capacitor-less AC-AC converter was developed based on the matrix converter and applied to a VSWPGS to control the power between the SRG and an electric system. The VSWPGS using the capacitor-less AC-AC converter and SRG was modeled with Matlab/Simulink. The model considered a mechanical model including a wind turbine, the SRG, the capacitor-less AC-AC converter, and a controller. The simulation results for wind velocity changes verified the performance of the proposed VSWPGS using the capacitor-less AC-AC converter and SRG.

This paper proposes simplified model order reduction for the fast dynamic simulation of electric motors. In this method, magnetic fields expressed by reduced variables are stored into memory and restored using interpolation in the current-state space to enable the properties of an electric motor to be computed fast. It is shown that the iron losses considering higher harmonics are computed accurately and effectively usingthe proposed method. Moreover, the proposed method is successfully coupled with flux-based circuit equations.

This study describes characteristics of an interior permanent magnet synchronous motor (IPMSM) using samarium cobalt (SmCo) magnets, operated at high ambient temperature up to 200℃. Although the SmCo magnet is not as strong as neodymium iron boron (NdFeB) magnets, it has less thermal demagnetization, and hence can operate at a wider temperature range and have better thermal stability than the NdFeB magnet. Two IPMSM prototypes using SmCo (LM-34SH) and NdFeB (N40-UH) magnets were built and tested at ambient temperatures up to 200℃. The test results demonstrated that the prototype IPMSM with SmCo magnets produced higher torque at more than 100℃ and had a wider high-efficiency range at 200℃ than that with NdFeB magnets, and hence it was experimentally verified that the IPMSM with SmCo has thermal motor performance superiority.

We are working on intelligent systems engineering with AI for the realization of intelligent social systems and intelligent medical and care support systems. This article introduces these efforts and cases.