Akihiko Nagakubo

ORCID: 0000-0002-0682-9237
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Robotic Locomotion and Control
  • Robot Manipulation and Learning
  • Advanced Sensor and Energy Harvesting Materials
  • Muscle activation and electromyography studies
  • Prosthetics and Rehabilitation Robotics
  • Modular Robots and Swarm Intelligence
  • Soft Robotics and Applications
  • Reinforcement Learning in Robotics
  • Electrical and Bioimpedance Tomography
  • Tactile and Sensory Interactions
  • Social Robot Interaction and HRI
  • Control and Dynamics of Mobile Robots
  • Interactive and Immersive Displays
  • Robotic Mechanisms and Dynamics
  • Genetic Neurodegenerative Diseases
  • Motor Control and Adaptation
  • Robotic Path Planning Algorithms
  • Embodied and Extended Cognition
  • Context-Aware Activity Recognition Systems
  • Hand Gesture Recognition Systems
  • Human Pose and Action Recognition
  • Machine Learning and ELM
  • Microfluidic and Bio-sensing Technologies
  • Analytical Chemistry and Sensors
  • IoT-based Smart Home Systems

National Institute of Advanced Industrial Science and Technology
2005-2023

Systems Research Institute
2011-2012

Intelligent Systems Research (United States)
2011-2012

The University of Tokyo
2012

Tokyo Institute of Technology
1991-2002

University of Tsukuba
2002

Electrotechnical Institute
2000

We present the design and realization of a conformable tactile sensor skin (patent pending). The is organized as network self-contained modules consisting tiny pressure-sensitive elements which communicate through serial bus. By adding or removing it possible to adjust area covered by well number (and density) elements. therefore highly modular thus intrinsically scalable. Moreover, because substrate on are mounted sufficiently pliable be folded stiff enough cut, freely distribute individual...

10.1109/robot.2006.1641896 article EN 2006-07-10

Jumping and landing movements are characterized by large instantaneous forces, short duration, a high uncertainty concerning take off points. Such characteristics make conventional types of control robot design inadequate. Here we present an approach to realize motor jumping which exploits the synergy between mechanical structure. Our experimental system is pneumatically actuated bipedal called "Mowgli". Mowgli's artificial musculoskeletal consists six McKibben pneumatic muscle actuators...

10.1109/robot.2007.363848 article EN Proceedings - IEEE International Conference on Robotics and Automation/Proceedings 2007-04-01

The configuration design for a wall-climbing robot which is capable of moving on diversified surfaces wall and has high payload capability, discussed, developed quadruped robot, NINJA-1, introduced. NINJA-1 composed (1) legs based 3D parallel link mechanism producing powerful driving force the surface wall, (2) conduit-wire-driven parallelogram to adjust posture ankles, (3) valve-regulated multiple sucker can provide suction even if there are grooves small differences in level wall. Finally,...

10.1109/icar.1991.240585 article EN 1991-01-01

The development of a wall climbing robot which is able to move over the surface "3-dimensional terrain", terrain including floor, wall, ceiling any kinds structures with agility and adaptability strongly demanded in many industries. conventional wall-climbing machines were far from fulfilling demand, authors thus have been developing four-terrain-adaptive legs basic mobilities already demonstrated by manufactured prototype model NINJA-I. However, as there are almost no examples quadrupedal...

10.1109/robot.1994.351225 article EN 2002-12-17

CPG (central pattern generator) and entrainment dynamics together form a promising framework for robust adaptive behavior generation high degree of freedom system in unstructured environment. This paper investigates its possibility the domain biped robotic locomotion. We extend previous work on 2D locomotion using neural oscillators to 3D, introducing many more degrees complexity control. While problem has been increased, we have simplified internal mechanism compared original work. Our...

10.1109/iros.1998.724601 article EN 2002-11-27

For robots to possess high dexterity and interact intelligently with objects, a deformable tactile sensor skin which can also sense the deformation is of great importance. However conventional sensors are not suitable for this purpose mostly due embedded wiring inside sensing area. To overcome problem, we have implemented EIT (electrical impedance tomography), technique based on inverse problem theory, estimate resistivity distribution within an electrical conductor by using only...

10.1109/robio.2007.4522352 article EN 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO) 2007-12-01

Recently, we have been studying various tactile distribution sensors based on electrical impedance tomography (EIT) which is a non-invasive technique to measure the resistance of conductive material only from boundary, and needs no wiring inside sensing area. In this paper, present newly developed structure pressure sensitive but stretch insensitive concept contact between (1)a network stretchable wave-like yarns with high (2)a sheet low resistance. Based structure, realized novel sensor...

10.1109/3dui.2009.4811210 article EN 2009-01-01

In this article, the design principle of leg driving mechanism to minimize energy loss and maximize output power is discussed. We will also introduce gait control methods implemented in our previous quadruped walking robots. Finally, we survey most prototype models

10.1109/mra.2009.932524 article EN IEEE Robotics & Automation Magazine 2009-06-01

Stretchable tactile distribution sensors are significantly important in realizing natural and rich interactions for next generation humanoids with soft smooth surfaces. We have developed such a sensor which has no wiring most of the sensing area is based on inverse problem technique - EIT. At first, we show ability to detect novel involving skin deformation as pinching rubbing. Furthermore, by using our newly conductive knitted fabric, demonstrate how can be extremely useful pressure...

10.1109/ichr.2007.4813864 article EN 2007-11-01

Deep imitation learning is promising for robot manipulation because it only requires demonstration samples. In this study, deep applied to tasks that require force feedback. However, existing methods have deficiencies; bilateral teleoperation a complex control scheme and expensive, kinesthetic teaching suffers from visual distractions human intervention. This research proposes new master-to-robot (M2R) policy transfer system does not robots feedback-based tasks. The directly demonstrates...

10.1109/lra.2023.3262423 article EN IEEE Robotics and Automation Letters 2023-03-28

To understand the human intelligence involved physical manipulation of objects, detailed assessment states contact between surface hand and objects is necessary. This paper proposes a tactile sensing glove based on branch-shaped sensor sheet that enables pressure measurement at 1052 points without critically interfering with movements joints hand. The change in joint moveability by wearing evaluated an experiment, result shows effectiveness proposed design. Finally, as preliminary result,...

10.1109/humanoids.2011.6100898 article EN 2011-10-01

10.1016/s0921-8890(01)00156-7 article EN Robotics and Autonomous Systems 2001-11-01

The configurational design for a wall-climbing robot which is capable of moving on diversified surfaces wall and has high payload capability, discussed, developed quadruped robot, NINJA-1, introduced. NINJA-1 composed 1) legs based 3D parallel link mechanism producing powerful driving force the surface wall, 2) CP (Conduit-wire-driven Parallelogram) to adjust posture ankles, 3) VM (Valve-regulated Multiple) sucker can provide suction even if there are grooves small differences in level wall....

10.1299/kikaic.57.2679 article EN TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C 1991-01-01

This paper considers the humanoid research as an approach to understanding and realizing complex real world interactions among robot, environment, human. As a first step towards extracting common principle over three term interactions, concept of action oriented control has been investigated with simulation example. The interaction view casts unique constraints on design humanoid, such whole body, smooth shape non-functional-modular design. A brief description ongoing ETL-humanoid which...

10.1109/iros.1997.655104 article EN 2002-11-23

CPG (central pattern generator) and entrainment dynamics together form a promising framework for robust adaptive behavior generation high degree of freedom system in unstructured environment. This paper investigates its possibility the domain biped robotic locomotion. We extend previous work on 2D locomotion using neural oscillators to 3D, introducing many more degrees complexity control. While problem has been increased, we have simplified internal mechanism compared original work. Our...

10.7210/jrsj.18.87 article EN Journal of the Robotics Society of Japan 2000-01-01

Presents the hardware of upper body ETL-Humanoid system. It has 24 degrees freedom with high performance actuators, which gives system mobility, such as is required for pushups, lifting itself or another person. The mechanical design able to achieve human proportions, smooth outer shape, and light weight at a very level, considering motor embedded inside. Custom compact electronics AC servo control fast network measurement was also developed. still evolving, being used working testbed...

10.1109/iros.2000.895198 article EN 2002-11-08

Whole-body dynamic actions under various contacts with the environment will be very important for future humanoid robots to support human tasks in unstructured environments. Such skills are difficult realize using standard motion control methodology based on asymptotic convergence successive desired states. An alternative approach would exploit passive dynamics of body constrained motion, and navigate through multiple by imposing least order robustly reach goal state. As a first example such...

10.1142/s021984360400023x article EN International Journal of Humanoid Robotics 2004-09-01

To understand the principles of human hand dexterity, investigation dynamic contact control by is essential. This paper proposes a novel high-density tactile sensing glove lined with sensor sheets, which were flexible printed circuit boards embedded 1-[mm]-sized pressure sensitive elements at 1,052 points per hand. The sheets tree-shaped, especially designed to accommodate deformations skin in order minimize interference wearer's natural finger movements. design also permitted seamless...

10.7210/jrsj.30.711 article EN Journal of the Robotics Society of Japan 2012-01-01

Abstract This paper presents the final stage of development a humanoid system, ETL-Humanoid. It is full-scale system with 46 d.o.f., height and weight an average Japanese person. was designed as experimental platform to explore general principle controls complex embodied systems. The complete mechanical configuration low-level network-based control will be presented. possesses properties compactness, modularity light in weight. high performance, back-drivable compliant, allowing possibility...

10.1163/156855303321165105 article EN Advanced Robotics 2003-01-01

This paper presents the final stage of development humanoid system, ETL-Humanoid. It is full-scale system with 46 degrees freedom, height and weight an average Japanese person. was designed as experimental platform, to explore general principle controls complex embodied systems. The complete will be presented; mechanical configuration low-level network-based control also presented. possesses properties compactness, modularity light in weight. high performance, backdrivable compliant,...

10.1109/iros.2001.976313 article EN 2002-11-13

Dexterous dynamic actions will be very important for future humanoid robots to support human tasks in unstructured environments. Such skills are difficult achieve by the current standard control strategy based on asymptotic convergence successive desired states. An alternative approach would exploit natural interaction dynamics between body and environment, navigate through multiple imposing least order robustly reach goal state. As a first example of such strategy, we propose investigate...

10.7210/jrsj.23.706 article EN Journal of the Robotics Society of Japan 2005-01-01
Coming Soon ...