論文 - 田向　権

Low-cost computation for isolated sign language video recognition with multiple reservoir computing 査読有り

Syulistyo A.R., Tanaka Y., Pramanta D., Fuengfusin N., Tamukoh H.

Plos One   20 ( 7 JULY )   2025年07月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Sign language recognition (SLR) has the potential to bridge communication gaps and empower hearing-impaired communities. To ensure the portability and accessibility of the SLR system, its implementation on a portable, server-independent device becomes imperative. This approach facilitates usage in areas without internet connectivity, addressing the need for data privacy protection. Although deep neural network models are potent, their efficacy is hindered by computational constraints on edge devices. This study delves into reservoir computing (RC), which is renowned for its edge-friendly characteristics. Through leveraging RC, our objective is to craft a cost-effective SLR system optimized for operation on edge devices with limited resources. To enhance the recognition capabilities of RC, we introduce multiple reservoirs with distinct leak rates, extracting diverse features from input videos. Prior to feeding sign language videos into the RC, we employ preprocessing via MediaPipe. This step involves extracting the coordinates of the signer’s body and hand locations, referred to as keypoints, and normalizing their spatial positions. This combined approach, which incorporates keypoint extraction via MediaPipe and normalization during preprocessing, enhances the SLR system’s robustness against complex background effects and varying signer positions. Experimental results demonstrate that the integration of MediaPipe and multiple reservoirs yields competitive outcomes compared with deep recurrent neural and echo state networks and promises significantly lower training times. Our proposed MRC achieved accuracies of 60.35%, 84.65%, and 91.51% for the top-1, top-5, and top-10, respectively, on the WLASL100 dataset, outperforming the deep learning-based approaches Pose-TGCN and Pose-GRU. Furthermore, because of the RC characteristics, the training time was shortened to 52.7 s, compared with 20 h for I3D and the competitive inference time.

DOI： 10.1371/journal.pone.0322717

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105012039532&origin=inward

Inner Speech Recognition Using EEG Signals with Reservoir Computing 査読有り

Syulistyo Arie Rachmad, Tanaka Yuichiro, Tamukoh Hakaru

International Journal of Activity and Behavior Computing ( 九州工業大学ケアXDXセンター )   2025 ( 2 )   1 - 19   2025年06月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Inner speech recognition using electroencephalography (EEG) signals holds significant potential for advancing brain-computer interface (BCI), particularly for individuals with speech impairments. However, decoding inner speech from EEG data remains a challenging task due to the nonlinear, high-dimensional and temporal dynamic nature of neural signals. In order to address the challenges, this study explores the application of reservoir computing (RC), with a particular focus on bidirectional RC, for the purpose of classifying inner speech from EEG signals. Contrary to unidirectional RC, which processes data in a single time direction, bidirectional RC captures both past and future temporal dependencies. This enhancement of the extraction of meaningful EEG features for classification is a significant contribution of this study. We evaluate the performance of both unidirectional and bidirectional RC architectures across a range of reservoir sizes (400, 500, and 600 units) to identify the most effective configuration for this task. Our results demonstrate that the bidirectional RC consistently outperforms the unidirectional RC in terms of accuracy and F1-score across all reservoir sizes, highlighting its superior ability to extract comprehensive temporal features from EEG data. The optimal performance is achieved by bidirectional RC with 600 reservoir units, yielding an accuracy of 18.94% and an F1-score of 19.02%, which surpasses all other configurations. In contrast, unidirectional RC with 600 units exhibits a lower accuracy of 17.22%. These findings underscore the potential of bidirectional RC in EEG signals classification for inner speech recognition, offering a promising direction for developing efficient BCI with low-cost computation.

DOI： 10.60401/ijabc.117

CiNii Research

AN INTEGRATIVE METHOD OF DETECTION AND LOCALIZATION OF SURROUNDING VEHICLES USING GPS AND RGB CAMERA DEVICE ATTACHMENTS 査読有り

Davaanyam J., Koga T., Al Aama O., Taniguchi T., Tamukoh H., Wagatsuma H.

Icic Express Letters   19 ( 6 )   627 - 635   2025年06月

　詳細を見る

記述言語：英語   掲載種別：研究論文（学術雑誌）

This study aims to enhance environmental monitoring in a self-driving bus project within an urban environment. Firstly, as previously mentioned, the positions of vehicles in traffic near the self-driving bus should be synchronized. While some researchers utilize LIDAR, RADAR, and camera technologies for environmental monitoring and mapping, these methods have high data requirements and are particularly challenging to implement in real-time operations. This approach will ensure the identification of vehicles using YOLOv9 and the collection of their coordinates. The Pixel Map Match and Filter method will be used to minimize GPS data errors, and camera data will be combined with GPS data based on GPS data interpolation to determine the factors affecting vehicle interaction space. This paper presents the results of this integration, demonstrating improved accuracy and reliability compared to traditional systems, and proposes a technology that can contribute to the development of self-driving vehicles.

DOI： 10.24507/icicel.19.06.627

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105007464500&origin=inward

カーボンナノチューブスポンジ構造を用いた物理リザバー素子とそのインセンサ演算応用 査読有り

君塚 紘喜, アズハリ サマン, 徳野 将士, カラチャリ アハメト, 宇佐美 雄生, 池本 周平, 田向 権, 田中 啓文

計測と制御 ( 公益社団法人 計測自動制御学会 )   64 ( 5 )   327 - 330   2025年05月

　詳細を見る

記述言語：日本語   掲載種別：研究論文（学術雑誌）

DOI： 10.11499/sicejl.64.327

CiNii Research

その他リンク： http://id.ndl.go.jp/bib/034186495

Haptic In-Sensor Computing Device Based on CNT/PDMS Nanocomposite Physical Reservoir 査読有り

Kimizuka K., Azhari S., Tokuno S., Karacali A., Usami Y., Ikemoto S., Tamukoh H., Tanaka H.

Advanced Intelligent Systems   7 ( 3 )   2025年03月

　詳細を見る

記述言語：英語   掲載種別：研究論文（学術雑誌）

The importance of haptic in-sensor computing devices has been increasing. Herein, a haptic sensor with a hierarchical structure is successfully fabricated via the sacrificial template method, using carbon nanotube-polydimethylsiloxane (CNT-PDMS) nanocomposites for in-sensor computing applications. The CNT-PDMS nanocomposite sensors, with different sensitivities, are obtained by varying the amount of CNTs. The input stimuli are transformed into higher-dimensional information, enabling a new path for the CNT-PDMS nanocomposite application, which is implemented on a robotic hand as an in-sensor computing device by applying a reservoir computing paradigm. The nonlinear output data obtained from the sensors are trained using linear regression and used to classify nine different objects used in everyday life with an object recognition accuracy of >80% for each object. This approach can enable tactile sensation in robots while reducing the computational cost.

DOI： 10.1002/aisy.202400640

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=105001063382&origin=inward

A Data Format Integration of Open-Street-Map and Lanelet2 Toward the Ontology Framework for Safety Automated driving systems 査読有り

Al Aama Obada, Koga Takahiro, Taniguchi Tomoki, Jargal Davaanyam, Oishi Junya, Nemoto Shigeru, Mizushina Wataru, Hirao Kazuki, Tamukoh Hakaru, Wagatsuma Hiroaki

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   808 - 814   2025年02月

　詳細を見る

記述言語：英語   掲載種別：研究論文（学術雑誌）

This study proposes a framework that integrates OpenStreetMap (OSM) data with ontology-based systems to enhance automated driving. OSM provides static geographical data, while the Lanelet2 mapping framework incorporates lanelevel road information and topological relationships. This combination enables advanced testing of vehicle behavior in realistic environments. Ontology-based integration offers semantic representations of road elements and traffic rules, supporting the modeling of complex driving scenarios. By structuring spatial and semantic data, this approach ensures accurate simulation and testing, facilitating applications such as traffic analysis, route optimization, and automated driving decision-making. The framework enhances scalability, precision, and safety in autonomous vehicle development, enabling more effective testing and validation. This integration supports comprehensive, context-aware simulations that improve vehicle response to real-world driving conditions.

DOI： 10.5954/icarob.2025.gs1-3

CiNii Research

Robotic Grasping of Common Objects: Focusing on Edge Detection for Improved Handling 査読有り

Shiba Tomoya, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   487 - 490   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Grasping objects like plates and cups poses unique challenges for robots because of their irregular shapes and the difficulty of finding reliable grasp points. Traditional approaches often attempt to grasp the object at its center, but this strategy tends to fail for items like plates or cups, whose shapes deviate from simple forms like cubes or spheres. To address this issue, we propose a new method that utilizes AI-powered image analysis to identify the best edges for grasping. Through experiments conducted with a home service robot and a set of YCB objects, we evaluated the effectiveness of our approach compared to conventional methods. The results revealed a significant improvement in the success rate, particularly for objects with prominent edges, such as cups.

DOI： 10.5954/icarob.2025.os17-5

CiNii Research

Proposal of a Grasp Verification Method Utilizing Background Subtraction and Depth Information 査読有り

Terashima Ryo, Yano Yuga, Arimura Koshun, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   491 - 494   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Commonly grasp verification approach involves using the opening width of the robot's gripper. However, methods based on the opening width of the gripper may not apply to slender objects. In this study, we propose a grasp verification method using background subtraction. Our proposed method uses depth information to mask the background, isolating only the images of the gripper and the grasped object. Subsequently, a difference image is created by comparing the current image with the pre-grasp state, and the grasp state is detected based on the magnitude of the observed changes. The method minimizes environmental influences by masking the background, enabling highly accurate grasp verification even for complex objects. Through experiments, we validate the effectiveness of the proposed method.

DOI： 10.5954/icarob.2025.os17-6

CiNii Research

Prediction of Timing and Amount of Houseplants Watering by an Echo State Network on Jetson 査読有り

Yoshimura Wataru, Arimura Koshun, Kobayashi Ryohei, Mizutani Akinobu, Fujino Tomoaki, Tanaka Yuichiro, Sudo Tomomi, Ishizuka Naoto, Ito Keitaro, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   529 - 534   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Cultivating of houseplants in biophilic designed spaces requires appropriate timing and amount of watering. However, determining them is challenging, as fluctuations in ambient temperature can influence these factors. We develop a system capable of predicting ambient temperature changes and determining the appropriate timing and amount of watering. The system acquires ambient data using sensors connected to a Jetson Nano and processes the data using a neural network for the prediction and determination. We adopt an echo state network, a lightweight neural network, enabling a power-efficient system capable of running on edge devices. Additionally, we implement a function to notify the user of the timing and amount of watering via a chat service whenever the soil moisture content drops below a predefined threshold.

DOI： 10.5954/icarob.2025.os18-4

CiNii Research

Optimizing Object Placement for Human Support Robots Using a Two-dimensional Irregular Packing Algorithm for Efficient Tray Storage 査読有り

Buttawong Natee, Isomoto Kosei, Yamao Kosei, Fuengfusin Ninnart, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   471 - 475   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

Human support robots (HSR) are robots that assist humans in their daily tasks. Their main application is tidying up, which involves detecting objects, determining appropriate placement locations, and organizing them. This study focuses on tidying up tray storage. Determining suitable storage positions is essential for storing objects in the tray. If the robot store objects in fixed predetermined locations, it can lead to inefficient use of storage space, and in the worst case, objects might collide and overflow from the tray. To address this limitation, we propose a 2-dimensional irregular packing algorithm utilizing an object mask method to calculate the best placement location. This study evaluates the proposed packing algorithm against the standard method to determine which approach is more effective in HSR applications.

DOI： 10.5954/icarob.2025.os17-1

CiNii Research

Integrating Advance Speech Recognition and Human Attribute Detection for Enhanced Receptionist Tasks in RoboCup@Home 査読有り

Arimura Koshun, Yano Yuga, Kawabata Takuya, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   480 - 483   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

RoboCup@Home is held to integrate service robots into society. It includes a task called "Receptionist" that evaluates Human-Robot Interaction. In this task, a robot must ask guests for their names and favorite drinks and guide them to available seats. Additionally, the robot must introduce the guest's features, such as their clothing, to others. We developed a system integrating speech recognition and human attribute detection to achieve these functions. The robot can determine which seat a person is sitting in by detecting the person's skeletal coordinates. Additionally, the robot can identify individuals by recognizing human attributes. To verify the effectiveness of the developed system, we participated in the Receptionist task at RoboCup@Home 2024. We won first place in our league and demonstrated the effectiveness of our system.

DOI： 10.5954/icarob.2025.os17-3

CiNii Research

Grasp Point Estimation using Simulator-Generated Datasets Including Pose Information 査読有り

Maruno Ryoga, Shiba Tomoya, Yamaguchi Naoki, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   495 - 498   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

We develop a system that automatically generates training datasets for object recognition models using a simulator. In this study, we incorporate pose information into the dataset. Using this system, we develop a method for estimating grasp points for objects that are difficult for robots to grasp, selecting a toy airplane as the target object. Three specific points are assigned to the object: the front, center, and back. In the grasp point estimation process, the center point is designated as the grasp point. The robot's arm achieves an appropriate grasp by moving perpendicularly to the line connecting the front and back points toward this grasp point.

DOI： 10.5954/icarob.2025.os17-7

CiNii Research

Efficient Object Detection with Color-Based Point Prompts for Densely Packed Scenarios in WRS FCSC 2024 査読有り

Yamaguchi Naoki, Shiba Tomoya, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   509 - 512   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper introduces a novel object detection method designed for densely packed environments, such as those encountered in the World Robot Summit Future Convenience Store Contest (FCSC) 2024. Our system leverages color-based point prompts in conjunction with Segment Anything (SAM) 2 to achieve precise object segmentation and grasp point estimation, specifically targeting scenarios where objects like rice ball cluster tightly within containers. Unlike traditional methods that depend on pre-defined grasping strategies susceptible to positional drift in mobile robots, our approach dynamically identifies and isolates objects without requiring extensive retraining inherent to CNN or Transformer models. We conducted comprehensive experiments comparing our method against SAM 2 and Grounding DINO using a dataset of 10 test images containing 202 rice balls. Additionally, we deployed the system on a Toyota Human Support Robot during the FCSC Stock Task to assess real-world performance metrics. Results demonstrate that our method achieves higher detection accuracy and operational efficiency, validating its potential for autonomous retail applications.

DOI： 10.5954/icarob.2025.os17-10

CiNii Research

Developing a Sound-Based Method to Synchronize Multiple Videos Recorded by Multiple Sound Sources 査読有り

Jargal Davaanyam, Kato Rena, Taniguchi Tomoki, Shibata Kosei, Koga Takahiro, Al Aama Obada, Tamukoh Hakaru, Wagatsuma Hiroaki

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   815 - 820   2025年02月

　詳細を見る

記述言語：英語   掲載種別：研究論文（学術雑誌）

The surround environmental monitoring from a vehicle on the road is important for the driving behavior analysis as well as the monitoring of driver's operations. A light weight mobile camera is useful for the record of multiple directions from the vehicle simultaneously. However, the timing synchronization is an important issue need to be solved. In this purpose, we proposed the sound-based method to synchronize different videos recorded with environmental sounds. In this task, the extraction of common sound features and amplifying of the features are necessary to superimpose those sound profiles to find consisting time points. In the validation the effectiveness, we used recoded videos from the bus driven by an expert driver.

DOI： 10.5954/icarob.2025.gs1-4

CiNii Research

Classification of Human Activity by Spiking Neural Networks using Event-based Vision Sensors 査読有り

Rakheja Aryan, Saini Rohan, Toyoda Ryuta, Tanaka Yuichiro, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   484 - 486   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

This paper presents a human action classification system using a spiking neural network (SNN) and an event-based vision sensor (EVS). The EVS captures asynchronous data with high temporal resolution, wide dynamic range, and motion blur resistance. SNNs, inspired by biological neurons, process this data event-driven, ensuring energy efficiency, low latency, and scalability. Using a custom EVS dataset of 600 videos across four action types and the optical flow for feature extraction, the system achieved 93% accuracy, offering an efficient solution for action recognition.

DOI： 10.5954/icarob.2025.os17-4

CiNii Research

Classification of Human Activity by Event-based Vision Sensors using Echo State Networks 査読有り

Saini Rohan, Rakheja Aryan, Toyoda Ryuta, Tanaka Yuichiro, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   476 - 479   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

We propose a system for human activity recognition using an event-based vision sensor (EVS) with echo state networks (ESNs). Conventional cameras are susceptible to motion blur and require computationally intensive methods, whereas EVS provides no motion blur and low latency. Our research aims to enable accurate recognition of human activities by using energy-efficient methods. Therefore, we adopt ESNs, which require low computational costs, for the classifier. Additionally, we use feature extraction algorithms such as optical flow and histogram of gradients to improve accuracy. We used an EVS activity recognition dataset created by us containing six human activities and 600 videos. The results showed that our hybrid approach outperformed several techniques. We achieved 89% accuracy when trained with ridge regression.

DOI： 10.5954/icarob.2025.os17-2

CiNii Research

Basic Research on the Development of Space Standards for the Use of Service Robots in Housing Using the Urban Renaissance Agency's Housing Complex 査読有り

Matsuoka Ren, Nonoshita Kanon, Ishizuka Naoto, Kobayashi Ryohei, Mizutani Akinobu, Tamukoh Hakaru, Tanaka Hirofumi

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   519 - 523   2025年02月

　詳細を見る

記述言語：英語   掲載種別：研究論文（学術雑誌）

This study conducts basic research on developing space standards for integrating home service robots in residences, using a room in an housing complex from Urban Renaissance Agency as a model. The study compares these spaces with existing robot mobility standards to identify issues. The robot-friendly level of the room was evaluated using RFA standard. Results showed that most rooms fell into Level C, the lowest of three levels, in areas such as "fixture width" and "steps." Furthermore, when operating two types of home service robots, the study uncovered unique robot behavior issues in the housing that were not anticipated by RFA standard.

DOI： 10.5954/icarob.2025.os18-2

CiNii Research

Application of AI Robot Technology for Biophilic Design 査読有り

Manabe Kairi, Miyazono Ryo, Ito Keitaro, Sudo Tomomi, Ishizuka Naoto, Mizutani Akinobu, Tamukoh Hakaru, Tanaka Yuichiro, Ueda Etsushi, Anamizu Yuki, Furuya Leon, Tamai Honoka, Nakano Saya, Tanaka Hirofumi

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   513 - 518   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

To enhance human health and well-being, Biophilic design has been increasingly recognized in recent years. This design is characterized by the integration of natural elements such as plants, nature light, and water into spaces. However, challenges are posed by the maintenance of live plants, as their decline can be caused by insufficient environmental conditions. The aim of this study is to propose a robotic system of autonomously relocate houseplants to environments optimized based on sensor data, including light, temperature, and humidity. Through the integration of AI robotics with ecological design principles, it is aimed to enhance sustainability and redefine the relationship between humans, nature, and technology, fostering a harmonious interaction among "robots, nature, and humans.".

DOI： 10.5954/icarob.2025.os18-1

CiNii Research

An Exhibition Environment with 2D Markers for Guide Robot 査読有り

Mizutani Akinobu, Hattori Yui, Ishizuka Naoto, Tanaka Yuichiro, Tanaka Hirofumi, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   524 - 528   2025年02月

　詳細を見る

担当区分：責任著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

The guide robot in an exhibition environment is expected to entertain visitors and reduce the maintenance cost of updating the robot database according to the changes in exhibition contents. The exhibition comprises printed panels for humans and a 2D marker printed with ultraviolet ink for robots. 2D markers are attached to the bottom of the exhibition furniture to make the markers invisible to visitors. The exhibition point may change depending on the contents of the exhibition. By searching for 2D markers autonomously in an exhibition space, the robot can update its internal database. This enables us to keep both the printed panel and the robot database without changing the robot database by users. The experiment is conducted in an exhibition environment, and the success rate in finding the exhibition point is evaluated.

DOI： 10.5954/icarob.2025.os18-3

CiNii Research

A feasibility study of generative AI applications using EV-3 Robots at the Kyushu Institute of Information Sciences 査読有り

Pramanta Dinda, Tamukoh Hakaru

人工生命とロボットに関する国際会議予稿集 ( 株式会社ALife Robotics )   30   499 - 503   2025年02月

　詳細を見る

担当区分：最終著者   記述言語：英語   掲載種別：研究論文（学術雑誌）

The advent of generative artificial intelligence (Gen AI) using large language models (LLMs) has brought about a transformation in the fields of education and home robotics. This study examines how students at the Kyushu Institute of Information Sciences perceive and utilize ChatGPT, with a particular on their familiarity, ethical considerations, and trust (FET) concerns. To further investigate this, we developed an educational EV-3 robot powered by ChatGPT and simulate the voice command using socket connections. By combining surveys with hands-on experiments, we uncovered the strengths and limitations of both the educational and home robotics roles. Our findings highlight the importance of FET, and 88% of respondents have no issues with implementing such technologies in the future.

DOI： 10.5954/icarob.2025.os17-8

CiNii Research