Nobuyasu Tomokuni

Ubiquitous Localization and Mapping for Robots with Ambient Intelligence

In this paper, a novel approach of knowledge management for the space/location perceiving capacity of robots is proposed. First, the ubiquitous function services composed of smart object, smart logic, and smart discovery service are proposed in order to distribute knowledge flexibly and reliably to changing environment and also for a robot to invoke and merge the distributed knowledge more freely. Next, physical and virtual spaces are merged by RFID tags. Through this, it is shown that ambient intelligence is realized and the space localization and mapping problem of robots can be more easily solved. Finally, the experiment based on an example scenario is carried out to verify the proposed method in the informative space named u-RT space which has two kinds of RFID tags, a virtual space with ubiquitous function services realized by Web services, and a networked robot system which works in these spaces

Development of ubiquitous mobile manipulator system with RT-Middleware

In this paper, the ubiquitous mobile manipulator system, which could realize the concept of ubiquitous robotics, is proposed. The ubiquitous robot might be able to give several services in human environments, beside classical robots have many problems to realize services, because of the complexity of environments and the amount of information. The experiment shows the validness of this system configuration. Our robots have 7DOF manipulator, mobile, gripper, tag reader and so on. And, all of systems are integrated on RT-middleware.

Ubiquitous Function Services Based Control for Robots with Ambient Intelligence

Ubiquitous computing brought about the paradigm change that robot technology transforms itself to technology for everything which includes robots and objects. Actually, these changes are continued to the various researches to link robots to the functions of objects which can be used by programmable resource on networks in the ubiquitous computing environment. In this paper, these inherent functions of objects are defined as ubiquitous functions. And ubiquitous function service with the service oriented architecture is proposed for integration and control of physical and virtual objects in the unified framework, which is composed of smart object service, smart logic service, and smart discovery service. Next, to provide flexible robustness for a robot in the everyday domain, the reproductive robust internal-loop compensator structure is proposed and the structural design method of smart logic service based on this structure is illustrated. Finally, em u-RT space for the proposed ubiquitous function services is introduced and the experiment based on an example scenario is carried out to verify the proposed design method of smart logic service

Distributed controller for cooperative joint torque control of compact humanoid robots

This paper describes the design issues of distributed control system for compact humanoid robots. First, structure of distributed servo controller is proposed which is connected to a real-time serial interface. This servo-controller system has torque control function by improving a servocontroller of a conventional radio-control servomechanism. By use of this proposed distributed controller, a compact humanoid robot with 22 degrees of freedom is made which has compliance control function. Experimental results give good evidence of the proposed distributed controller architecture

Proposal of ZMP Estimation Methods with Torque Information Estimated from Current Measurement and Its Application

In this paper, several models of approximating ZMP for humanoid are proposed without force sensors. Generally, ZMP is determined by using force sensor information. In this paper, this torque information is estimated by measuring motor-driving current. This paper discusses several estimation models whose results are compared by experiments. Furthermore, we examined effectiveness of proposed models. The results show that the proposed estimation methods are effective in semi-static condition. Specially, these proposed methods have advantages when a humanoid without a force sensor is controlled under change of gravity direction such as a slope condition.