研究者詳細 - 松野　隆幸

その他リンク： http://search.jamas.or.jp/link/ui/2020210402

Sea Docking by Dual-eye Pose Estimation with Optimized Genetic Algorithm Parameters 査読

Khin Nwe Lwin, Myo Myint, Naoki Mukada, Daiki Yamada, Takayuki Matsuno, Kazuhiro Saitou, Waichiro Godou, Tatsuya Sakamoto, Mamoru Minami

96 ( 2 ) 245 - 266 2019年11月

Experimental verification of turbidity tolerance of stereo-vision-based 3D pose estimation system 査読

Myo Myint, Khin Nwe Lwin, Naoki Mukada, Daiki Yamada, Takayuki Matsuno, Yuuichirou Toda, Saitou Kazuhiro, Mamoru Minami

24 ( 3 ) 756 - 779 2019年9月

Improving pose estimation accuracy and expanding of visible space of lighting 3D marker in turbid water

Horng-Yi Hsu, Naoki Mukada, Daiki Yamada, Khin Nwe Lwin, Myo Myint, Yuichiro Toda, Takayuki Matsuno, Keigo Watanabe, Mamoru Minami

2019 IEEE International Underwater Technology Symposium, UT 2019 - Proceedings 2019年4月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

Aiming at developing underwater battery recharging system, the author developed a docking system using stereo-vision-based visual servoing and a 3D marker. The 3D marker consists of red, green, blue spheres that do not emit the light which is named as a passive marker. Real-time relative pose (position and orientation) estimation was implemented utilizing the 3D model-based matching method and real-time multi-step genetic algorithm (RM-GA). Given the situation that the docking aims for battery recharging in the deep-sea bottom, the pitch-dark and turbid environment should be considered as an inevitable condition for battery recharging. In our previous works, the docking experiments were conducted in the actual sea, having verified the effectiveness of the proposed system using the passive 3D marker in the daytime environment with turbid water condition. Since lighting passive 3D marker by light from the vehicle in turbid water environment results in a situation that the images taken by video cameras set on the vehicle were looked wholly white, some new idea seems to be required. To overcome this difficulty, the newly lighting 3D marker (active 3D maker) has LEDs inside was introduced in the previous work. The main objective of this study is to check the feasibility area of the proposed system for the docking application, comparison of recognition performance using the active and passive 3D marker that was conducted in the simulated pool with the turbid water is focused. And then, the experiment using the active 3D marker in the actual sea has been performed. The experimental results have confirmed that the new active 3D marker with no-lighting from the vehicle could be recognizable in dark and turbid environment than the passive 3D marker with the lighting from the vehicle.

DOI： 10.1109/UT.2019.8734353

Improvement of 3D Pose Estimation Abilities by Light-Emitting-3D Marker for AUV Docking

Kohei Yamashita, Hsu Horng Yi, Daiki Yamada, Naoki Mukada, Khin New Lwin, Myo Myint, Yuichiro Toda, Takayuki Matsuno, Mamoru Minami

OCEANS 2018 MTS/IEEE Charleston, OCEAN 2018 2019年1月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

Disturbances of turbidity and low illuminance are problems in real sea areas when recognizing objects with cameras. Therefore, the recognition target was made to emit light so that it can be recognized correctly even in that environment. However, a suitable light intensity of the target was not decided and it is obvious that recognition results was changed by light intensity of the target. This paper presents the analysis of recognition accuracy of the Real-time 3D estimation system by changing the current value of each color LED (red, green, blue) under turbid and low illuminance. Recognition experiments were conducted at the distance 600 [mm] between the ROV and 3D marker. The turbidity level was set constant value. The current value was changing from 0 [mA] to 16 [mA] for each LED individually. The best current for each LED was optimized by the fitness value and estimation value of position and orientation. The results showed that the recognition accuracy of the proposed system was improved by using optimized lighting intensity.

DOI： 10.1109/OCEANS.2018.8604867

Applying Lighting Marker and Stereo-vision to V-shaped-thruster Vehicle for AUV Deep Sea Docking

Yoshiki Kanda, Myo Myint, Naoki Mukada, Daiki Yamada, Khin Lwin, Takayuki Matsuno, Yuichiro Toda, Mamoru Minami

OCEANS 2018 MTS/IEEE Charleston, OCEAN 2018 2019年1月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

A stereo-vision-based system of autonomous under-water vehicles (AUVs) for sea-bottom docking that enables for battery recharging to extend persistence time of underwater operation has been developed. This paper presents the docking experiment using a developed V-shaped-thruster typed under-water vehicle. A real-time 3 D pose (position and orientation) estimation method using a real-time multi-step genetic algorithm (RM-GA) has been proposed by the authors in previous works and used for docking based on 3D recognition a s a feedback pose information in real-time, named as 3D Move on Sensing (3D-MoS). Sea docking experiment results have confirmed the functionality and practicality of proposed docking approach using a hovering typed ROV in previous works. Since the hovering typed underwater vehicles are limited in mobilities concerning speed and operational space, verification of the 3D-MoS system using underwater vehicle that has more mobility deem to be meaningful direction for vision-based docking system to expand the utility value of AUVs. Therefore, in this study, control system for a new V-shaped-thruster typed vehicle is developed and docking experiment is conducted. This paper presents the development of the hardware design of V-shaped-thruster typed underwater vehicle and improvement of controlling with consideration of coupled configuration of thrusters.

DOI： 10.1109/OCEANS.2018.8604730

Development of Dual-eyes Docking System for AUV with Lighting 3D Marker

Sho Nakamura, Daiki Yamada, Naoki Mukada, Myo Myint, Khin New Lwin, Takayuki Matsuno, Yuichiro Toda, Mamoru Minami

OCEANS 2018 MTS/IEEE Charleston, OCEAN 2018 2019年1月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

This paper presents that our proposed docking system could apply for a new Remotely Operated Vehicle(ROV). The authors have proposed a 3D-perception based Move on Sensing (3D-MoS) system using a new 3D position and orientation (pose) estimation method with dual-eye camera that exploits the parallactic nature that enables reliable 3D pose estimation in real-time, named as 'Real-time Multi-step Genetic Algorithm (RM-GA).' We confirmed ROV have conducted docking that assumes charging battery under water by the system, having shown it effective. As a next step, docking experiment using the new ROV were conducted to verify the proposed system apply for new ROV. But, the new ROV is different from a ROV that have used previous experiments, on the point of thruster setting structure. Then, a new control system using Jacobian that shows relationship voltage and velocity was constructed. After it is confirmed the system is effective, docking experiments in the pool have been conducted. In this report, the structure and result of experiments are reported in detail.

DOI： 10.1109/OCEANS.2018.8604527

Region of interest growing neural gas for real-time point cloud processing

Yuichiro Toda, Xiang Li, Takayuki Matsuno, Mamoru Minami

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 11742 82 - 91 2019年

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Springer Verlag

This paper proposes a real-time topological structure learning method based on concentrated/distributed sensing for a 2D/3D point cloud. First of all, we explain a modified Growing Neural Gas with Utility (GNG-U2) that can learn the topological structure of 3D space environment and color information simultaneously by using a weight vector. Next, we propose a Region Of Interest Growing Neural Gas (ROI-GNG) for realizing concentrated/distributed sensing in real-time. In ROI-GNG, the discount rates of the accumulated error and utility value are variable according to the situation. We show experimental results of the proposed method and discuss the effectiveness of the proposed method.

DOI： 10.1007/978-3-030-27535-8_8

Robustness of 3D pose estimation against turbidity using dual-eye cameras and active/lighting 3D marker for visual-servoing based AUV

Khin Nwe Lwin, Myo Myint, Naoki Mukada, Daiki Yamada, Takayuki Matsuno, Mamoru Minami

2018 OCEANS - MTS/IEEE Kobe Techno-Oceans, OCEANS - Kobe 2018 2018年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

This paper presents the 3D pose estimation against turbidity under dark environment by using dual-eye cameras and an active —meaning light emitting— 3D marker for visual-servoing based underwater vehicle. The authors have proposed a 3D-perception based move on sensing (3D-MoS) system using a 3D position and orientation (pose) estimation method with dual-eye cameras that exploits the parallactic nature that enables reliable 3D pose estimation in real-time, named as "Real-time Multi-step Genetic Algorithm (RM-GA)." The active/lighting 3D marker was designed and constructed to improve the 3D pose estimation especially in turbidity and low illumination. In real-time pose estimation, not only recognition but also robustness are important. This paper focus on the robustness of 3D pose recognition performance using dual-eye cameras and 3D marker against turbidity. The experimental results have confirmed that the effectiveness and robustness of the proposed system for the real-time 3D pose estimation under turbidity and night condition.

DOI： 10.1109/OCEANSKOBE.2018.8559085

Docking experiment in dark environments using active/lighting marker and HSV correlation

Daiki Yamada, Naoki Mukada, Myo Myint, Khin Nwe Lwin, Takayuki Matsuno, Mamoru Minami

2018 OCEANS - MTS/IEEE Kobe Techno-Oceans, OCEANS - Kobe 2018 2018年12月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

Recharging ability with underwater docking function would be a first primal step conducted to enable the AUV to operate independently of a surface vessel for extended periods. Therefore, the role of docking operation came in picture not only for battery recharging application but also other novel applications such as sleeping under mother ship, or new mission up and down loading. Moreover, docking capacity can be extended to provide navigation for other underwater vehicles on the way of their mission too. However, there are many challenging issues in achieving these applications that request high accuracy and robustness against disturbances that are provided by the underwater environment. The most challenging and unavoidable problems in sensing sphere for sea operations are, we think, turbidity and light changing. Turbidity is denned as cloudiness in a liquid caused by the presence of suspended particles that scatter and absorb light. Since underwater battery recharging are supposed as a first step to realize a full autonomous/intelligent robot, the deep-sea docking experiments cannot avoid turbidity and low light environment. In previous studies, we had conducted sea docking experiments using a passive (not lighting) marker and image-evaluation function based on only hue information, limiting its operational environment in lower turbid sea with sunshine. Whereas in this study, to improve our system removing above defects, we newly designed an active - light emitting - 3D marker and a fitness function determined by HSV color components to improve the performance of the system especially in a more turbid environment. The advantage of using an active 3D marker and HSV-evaluated function is to be thought as being tolerable and seeable despite clipped whites and scattered light on the camera images caused by turbidity. Additionally, we conducted the docking experiments to verify the robustness of the proposed approach against turbidity and compared recognition results between the previous method and the improved method.

DOI： 10.1109/OCEANSKOBE.2018.8559303

Zerobot®: A Remote-controlled Robot for Needle Insertion in CT-guided Interventional Radiology Developed at Okayama University.

Takao Hiraki, Tetsushi Kamegawa, Takayuki Matsuno, Toshiyuki Komaki, Jun Sakurai, Susumu Kanazawa

Acta medica Okayama 72 ( 6 ) 539 - 546 2018年12月

記述言語：英語出版者・発行元：OKAYAMA UNIV MED SCHOOL

Since 2012, we have been developing a remote-controlled robotic system (Zerobot®) for needle insertion during computed tomography (CT)-guided interventional procedures, such as ablation, biopsy, and drainage. The system was designed via a collaboration between the medical and engineering departments at Okayama University, including various risk control features. It consists of a robot with 6 degrees of freedom that is manipulated using an operation interface to perform needle insertions under CT-guidance. The procedure includes robot positioning, needle targeting, and needle insertion. Phantom experiments have indicated that robotic insertion is equivalent in accuracy to manual insertion, without physician radiation exposure. Animal experiments have revealed that robotic insertion of biopsy introducer needles and various ablation needles is safe and accurate in vivo. The first in vivo human trial, therefore, began in April 2018. After its completion, a larger clinical study will be conducted for commercialization of the robot. This robotic procedure has many potential advantages over a manual procedure: 1) decreased physician fatigue; 2) stable and accurate needle posture without tremor; 3) procedure automation; 4) less experience required for proficiency in needle insertion skills; 5) decreased variance in technical skills among physicians; and 6) increased likelihood of performing the procedure at remote hospitals (i.e., telemedicine).

DOI： 10.18926/AMO/56370

PubMed

String Untying Planning Based on Knot Theory and Proposal of Algorithms to Generate the Motion of a Manipulator

Takayuki Matsuno, Tomoya Shirakawa, Tomotoshi Watanabe, Mamoru Minami

Proceedings - IEEE International Conference on Robotics and Automation 180 - 186 2018年9月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

Recently, the demand to manipulate deformable objects such as a string and cloth by robots is growing. The reason is that it has the possibility of making our lives more convenient in many domains. The manipulation of deformable objects, however, is more difficult than that of rigid objects, because deformable objects have diversity of shape and behavior. Therefore, our research group has been focusing on the string shape operation. This paper describes planning method of string untying operation based on knot theory and algorithms to generate the motion of a manipulator. The novel contribution of our planning method is automatic selection of optimal shape operation based on cost function. At final, the results of string untying experiments are reported.

DOI： 10.1109/ICRA.2018.8460477

Docking at pool and sea by using active marker in turbid and day/night environment

Khin Nwe Lwin, Naoki Mukada, Myo Myint, Daiki Yamada, Mamoru Minami, Takayuki Matsuno, Kazuhiro Saitou, Waichiro Godou

ARTIFICIAL LIFE AND ROBOTICS 23 ( 3 ) 409 - 419 2018年9月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SPRINGER

Nowadays, autonomous underwater vehicle (AUV) is playing an important role in human society in different applications such as inspection of underwater structures (dams, bridges). It has been desired to develop AUVs that can work in a sea with a long period of time for the purpose of retrieving methane hydrate, or rare metal, and so on. To achieve such AUVs, the automatic recharging capability of AUVs under the sea is indispensable and it requires AUVs to dock itself to recharging station autonomously. Therefore, we have developed a stereo-vision-based docking methodology for underwater battery recharging to enable the AUV to continue operations without returning surface vehicle for recharging. Since underwater battery recharging units are supposed to be installed in a deep sea, the deep-sea docking experiments cannot avoid turbidity and low-light environment. In this study, the proposed system with a newly designed activemeaning self-lighting3D marker has been developed to improve the visibility of the marker from an underwater vehicle, especially in turbid water. Experiments to verify the robustness of the proposed docking approach have been conducted in a simulated pool where the lighting conditions change from day to night. Furthermore, sea docking experiment has also been executed to verify the practicality of the active marker. The experimental results have confirmed the effectiveness of the proposed docking system against turbidity and illumination variation.

DOI： 10.1007/s10015-018-0442-1

Robotic Insertion of Various Ablation Needles Under Computed Tomography Guidance: Accuracy in Animal Experiments. 国際誌

Takao Hiraki, Takayuki Matsuno, Tetsushi Kamegawa, Toshiyuki Komaki, Jun Sakurai, Ryutaro Matsuura, Takuya Yamaguchi, Takanori Sasaki, Toshihiro Iguchi, Yusuke Matsui, Hideo Gobara, Susumu Kanazawa

European journal of radiology 105 162 - 167 2018年8月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：ELSEVIER IRELAND LTD

OBJECTIVE: To evaluate the accuracy of robotic insertion of various ablation needles at various locations under computed tomography (CT) guidance in swine. MATERIALS AND METHODS: The robot was used for CT-guided insertion of four ablation needles, namely a single internally cooled radiofrequency ablation (RFA) needle (Cool-tip), a multi-tined expandable RFA needle (LeVeen), a cryoablation needle (IceRod), and an internally cooled microwave ablation needle (Emprint). One author remotely operated the robot with the operation interface in order to orient and insert the needles under CT guidance. Five insertions of each type of ablation needle towards 1.0-mm targets in the liver, kidney, lung, and hip muscle were attempted on the plane of an axial CT image in six swine. Accuracy of needle insertion was evaluated as the three-dimensional length between the target centre and needle tip. The accuracy of needle insertion was compared according to the type of needle used and the location using one-way analysis of variance. RESULTS: The overall mean accuracy of all four needles in all four locations was 2.8 mm. The mean accuracy of insertion of the Cool-tip needle, LeVeen needle, IceRod needle, and Emprint needle was 2.8 mm, 3.1 mm, 2.5 mm, and 2.7 mm, respectively. The mean accuracy of insertion into the liver, kidney, lung, and hip muscle was 2.7 mm, 2.9 mm, 2.9 mm, and 2.5 mm, respectively. There was no significant difference in insertion accuracy among the needles (P = .38) or the locations (P = .53). CONCLUSION: Robotic insertion of various ablation needles under CT guidance was accurate regardless of type of needle or location in swine.

DOI： 10.1016/j.ejrad.2018.06.006

PubMed

Analyses of biped walking posture by dynamical-evaluating index

Keli Shen, Xiang Li, Hongzhi Tian, Takayuki Matsuno, Mamoru Minami

ARTIFICIAL LIFE AND ROBOTICS 23 ( 2 ) 261 - 270 2018年6月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SPRINGER

In this paper, biped walking posture and design are evaluated through dynamic reconfiguration manipulability shape index (DRMSI). DRMSI is the concept derived from dynamic manipulability and reconfiguration manipulability with remaining redundancy. DRMSI represents the ability of dynamical system of manipulators possessing shape changing acceleration in task space by normalized torque inputs, while the hand motion is assigned as the primary task. Besides, we use visual lifting approach to stabilize the walking and stop falling down. In this research, the primary task is to make the position of the head direct to the desired one as much as possible. And realizing the biped walking is the second task. This research indicates that proposed dynamical-evaluating index is effective in evaluating the biped walking motion and biped humanoid robot has the adjustable configuration to walk with higher flexibility. Flexibility represents the dynamical shape changeability of humanoid robot based on redundancy of the humanoid robot with the premise of the primary task given to keeping the head position high.

DOI： 10.1007/s10015-017-0415-9

Improved eye-vergence visual servoing system in longitudinal direction with RM-GA

Yejun Kou, Hongzhi Tian, Mamoru Minami, Takayuki Matsuno

ARTIFICIAL LIFE AND ROBOTICS 23 ( 1 ) 131 - 139 2018年3月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：SPRINGER

Visual servoing is a control method to manipulate the motion of the robot using visual information, which aims to realize "working while watching." However, the visual servoing towards moving target with hand-eye cameras fixed at hand is inevitably affected by hand dynamical oscillation. To overcome this defect of the hand-eye fixed camera system, an eye-vergence system has been put forward, where the pose of the cameras could be rotated to observe the target object. The visual servoing controllers of hand and eye-vergence are installed independently, so that it can observe the target object at the center of camera images through eye-vergence function. In this research, genetic algorithm (GA) is used as a pose tracking method, which is called "Real-Time Multi-step GA(RM-GA)," solves online optimization problems for 3D visual servoing. The performances of real-time object tracking using eye-vergence system and "RM-GA" method have been examined, and also the pose tracking accuracy has been verified.

DOI： 10.1007/s10015-017-0408-8