研究者業績

竹囲 年延

TOSHINOBU TAKEI

基本情報

所属
成蹊大学 理工学部 准教授 (Associate Professor)

J-GLOBAL ID
202301002868381931
researchmap会員ID
R000057872

論文

 36
  • Masahiro Inagawa, Keiichi Yoshizawa, Tomohito Kawabe, Toshinobu Takei
    J. Robotics Mechatronics 36(2) 320-333 2024年  
  • Masahiro Inagawa, Tomohito Kawabe, Toshinobu Takei, Keiji Nagatani
    ROBOMECH JOURNAL 10(1) 2023年7月  
    The construction industry faces a labor shortage problem, so construction vehicles need to be automated. For automation, a position estimation method is expected that is independent of the work environment and can accurately estimate the position of targets. This paper aims to develop a position estimation method for multiple construction vehicles using 3D LiDAR installed in a work environment. By focusing on the shape of construction vehicles, this method can estimate the location of construction vehicles in places where conventional methods cannot be used, such as valleys or roofs. Because the shape of the construction vehicle changes depending on the work equipment and steering operation, each joint angle was obtained, and the 3D model used for estimation was updated. As a result of the experiment, it was verified that the position and orientation of multiple construction vehicles can be estimated with an accuracy that satisfies the required accuracy.
  • Masahiro Inagawa, Tomohito Kawabe, Toshinobu Takei
    JOURNAL OF FIELD ROBOTICS 2023年5月  
    Localization methods for autonomous construction vehicles include real-time kinematic positioning through a global navigation satellite system (GNSS) and a scan matching with a light detection and ranging (LiDAR) attached to mobility. However, these conventional methods have low estimation accuracy when the vehicle's surroundings have few features and the vehicle is in the no-GNSS area. For the estimation in such areas, this paper proposes a localization method that can estimate by matching a 3D model of a construction vehicle with a point cloud obtained from 3D LiDARs installed in a work area. To realize the high-accuracy and high-speed processing of the localization, we propose remodeling using the predictive motion model (RM) algorithm to modify the 3D model in the registration process. In the experimental results on rough terrain, we confirmed that our method can estimate a vehicle's position and yaw angle with accuracies of 0.121 m and 0.016 rad, respectively. In addition, compared with the case without the RM algorithm, the construction vehicle's position and yaw angle accuracies improved up to 5 and 12 times, respectively.
  • Yuki Endo, Keisuke Yagi, Yoshikazu Mori, Toshinobu Takei, Hiromi Mochiyama
    ADVANCED ROBOTICS 37(8) 528-539 2023年4月  
    We propose a remote joint impedance estimation system called Tele-snap for a rehabilitation diagnosis under the COVID-19 pandemic. Dynamic resistance of the human joint is essential physical information reflecting the motor function. The resistance is assessed based on the touching sensation of the doctor (physiotherapist), but the pandemic restricts such an in-person manner. Our proposing system aims to provide this physical information quantified by the joint impedance for a diagnosis in the telerehabilitation context. The proposed system employs a compact impulsive perturbation generator called the snap motor and a marker-less motion capture technology called the OpenPose. The subsystem installed in the patient's place is then simplified remarkably, which consists of the wearable snap motor and Raspberry Pi with a built-in camera module. The proposed system can collect the dataset for impedance estimation through the examiner's teleoperation of the snap motor and camera via a virtual private network, with no need for the operation by the patient. We verify the proposed system through an in-person experiment and then demonstrate the remote impedance estimation scheme.
  • Tomohito Kawabe, Toshinobu Takei, Etsujiro Imanishi
    ADVANCED ROBOTICS 35(23) 1418-1437 2021年12月  
    This study introduces expedite the complete transfer of distributed gravel piles with an automated wheel loader. The wheel loader scoops the gravel and unloads it onto the bed of a truck. The total mileage for the repeated scooping and unloading work is reduced. The complicated optimal task is divided into three simple, organized subjects and different appropriate algorithms are used for each subject. A method is proposed that interactively selects the appropriate scooping points, unloading points, and appropriate path for the varying shape of the gravel pile until all gravel piles have been scooped up. To expedite the complete transfer of distributed gravel piles, the deep reinforcement learning model, trained via simulation, can be used in practice by implementing as part of the proposed methods. This paper describes these issues and the proposed methods. The performance of the actual application is demonstrated in terms of the calculation time and the feasibility of the simulations.

MISC

 84

共同研究・競争的資金等の研究課題

 7