Tomoaki Yoshikai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Tomoaki Yoshikai has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Masayuki Inaba
Department of Mechano-Informatics, The University of Tokyo, Tokyo, Japan
15 shared publications - Ikuo Mizuuchi
Tokyo University of Agriculture and Technology
9 shared publications Marika Hayashi
Univ. of Tokyo
6 shared publications- Tamaki Nishino
The University of Tokyo
5 shared publications - Yoshinao Sodeyama
The University of Tokyo
5 shared publications Yuto Nakanishi
The University of Tokyo
3 shared publications
Latest Publications
- 1 January 2012
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2012, _1A2-M01_1
1A2-M01 Achievement of Cooperative Climbing Motion by Reinforcement Behavior Learning Based on Sensory Integration of Multi-Humanoids(Humanoid(2))
It is thought that the cooperative action by humanoid robots will have an important meaning because they have flexible bodies which enable them to operate various motions. But self-conditions of the humanoid robots influence their actions, and it causes some troubles when they act, moreover when they cooperate. In this paper, we propose the reinforcement learning system based on integration of multi-humanoids senses. Constructing state space by using this sensory integration, robots learn the optimal movements. We expect that it makes them cover the weakest robot, solves the problems between the individual differences of robots, and enables them to achieve the tasks more flexibly. Using real humanoid robots, we certify the validity of the proposed technique.
- 1 January 2012
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2012, _1A2-M01_1
1A2-M01 Achievement of Cooperative Climbing Motion by Reinforcement Behavior Learning Based on Sensory Integration of Multi-Humanoids(Humanoid(2))
It is thought that the cooperative action by humanoid robots will have an important meaning because they have flexible bodies which enable them to operate various motions. But self-conditions of the humanoid robots influence their actions, and it causes some troubles when they act, moreover when they cooperate. In this paper, we propose the reinforcement learning system based on integration of multi-humanoids senses. Constructing state space by using this sensory integration, robots learn the optimal movements. We expect that it makes them cover the weakest robot, solves the problems between the individual differences of robots, and enables them to achieve the tasks more flexibly. Using real humanoid robots, we certify the validity of the proposed technique.
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _1P1-E10_1
1P1-E10 Impression Design and its Effect Based on an Appearance Feature Extraction of Robots Focused on "Flesh"
The purpose of this study is to obtain a novel useful description of a humanoid's appearance which is important for designing friendly and humanlike exterior. We focused on a biological word "flesh" and considered its design. Appearance features which can be called "skin surface" and "filling" were extracted and an robot exterior was designed considering these features. Effects of these features on oddness of the robot and tactile interaction were analyzed through questionnaire survey.
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _1P1-E10_1
1P1-E10 Impression Design and its Effect Based on an Appearance Feature Extraction of Robots Focused on "Flesh"
The purpose of this study is to obtain a novel useful description of a humanoid's appearance which is important for designing friendly and humanlike exterior. We focused on a biological word "flesh" and considered its design. Appearance features which can be called "skin surface" and "filling" were extracted and an robot exterior was designed considering these features. Effects of these features on oddness of the robot and tactile interaction were analyzed through questionnaire survey.
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _2A1-E05_1
2A1-E05 Stabilization of Camera View by Estimating the Camera Oscillation Based on a Local Correlation Matching Algorithm
Humanoid robots have high potentiality and 3D locomotion. Its walking motions make bigger camera oscillation than wheel robots. This can be a big problem when they try to get environmental information from cmaera view in walking. In this paper, to get more stable information of vision, we attempt to stabilize camera view based on calculation of 3D opticalflow.
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _2A1-E05_1
2A1-E05 Stabilization of Camera View by Estimating the Camera Oscillation Based on a Local Correlation Matching Algorithm
Humanoid robots have high potentiality and 3D locomotion. Its walking motions make bigger camera oscillation than wheel robots. This can be a big problem when they try to get environmental information from cmaera view in walking. In this paper, to get more stable information of vision, we attempt to stabilize camera view based on calculation of 3D opticalflow.
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _2A2-J01_1
2A2-J01 Multi-robot system with the ZigBee module
- 1 January 2009
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2009, _2A2-J01_1
2A2-J01 Multi-robot system with the ZigBee module
- 1 January 2008
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2008, _2P2-I11_1
2P2-I11 Psychological Evaluation of Soft Outer Coverings with Embedded Sensing and Expressing System for Representation of "Animacy"
- 1 January 2008
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2008, _2P2-I11_1
2P2-I11 Psychological Evaluation of Soft Outer Coverings with Embedded Sensing and Expressing System for Representation of "Animacy"
- 1 January 2007
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2007, _1A1-A05_1
1A1-A05 Acquisition of vision-based crank rotating behavior with spine for Musculo-skeletal humanoid Kotaro
Musculo-skeletal humanoid has many DOFs and flexible structure, so it can adapt to environment using hardware. To realize adaptation to environment, it is important for musculo-skeletal humanoid to modify its actions by using its own sensors because it is difficult to model for its complex structure. In this paper, by vision-based reinforcement learning, musculo-skeletal humanoid Kotaro aquires crank rotation behavior with flexible spine.
- 1 January 2007
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2007, _1A1-A05_1
1A1-A05 Acquisition of vision-based crank rotating behavior with spine for Musculo-skeletal humanoid Kotaro
Musculo-skeletal humanoid has many DOFs and flexible structure, so it can adapt to environment using hardware. To realize adaptation to environment, it is important for musculo-skeletal humanoid to modify its actions by using its own sensors because it is difficult to model for its complex structure. In this paper, by vision-based reinforcement learning, musculo-skeletal humanoid Kotaro aquires crank rotation behavior with flexible spine.
- 1 January 2007
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2007, _2A1-M04_1
2A1-M04 Simulation and Realization of Rolling-Over Motion with Soft Exterior Humanoid
With soft exterior and tactiles, humanoid robots will be more adaptive to contact with daily environment or with self body, and will have closer relation with surroundings. But, such close interaction causes complicated contacts, and rule of tactile feedback to motion can hardly be derived with analytic approach. This paper will show that a tactile feedback rule in a rolling-over motion (as an example of close interaction mentioned above) of the baby sized humanoid can be automatically acquired with simulation based evolutionary robotics approach. At last, the acquired feedback rule is tested using real robot.
- 1 January 2007
- Published by Japan Society of Mechanical Engineers in The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
- Vol. 2007, _2A1-M04_1
2A1-M04 Simulation and Realization of Rolling-Over Motion with Soft Exterior Humanoid
With soft exterior and tactiles, humanoid robots will be more adaptive to contact with daily environment or with self body, and will have closer relation with surroundings. But, such close interaction causes complicated contacts, and rule of tactile feedback to motion can hardly be derived with analytic approach. This paper will show that a tactile feedback rule in a rolling-over motion (as an example of close interaction mentioned above) of the baby sized humanoid can be automatically acquired with simulation based evolutionary robotics approach. At last, the acquired feedback rule is tested using real robot.