We propose the method to get three dimensional positional information by distributed LRF. When a robot moves, the positional information of the object is needed. LRF can measure distance correctly and acquire the positional information of around objects. It is efficient that multiple LRF are used and a sensor network is built. In this paper, we described our method of distributed LRF system in environment.

View full abstract

We propose a method that uses Monte Carlo localization to estimate the position and pose of robot on the lattice of RFID tags. Our experiments show that robots can localize their position more precisely than the interval of tags and also estimate their pose successfully by using the proposed method when two readers are placed in appropriate positions.

View full abstract

Wabot-House Research Laboratory is working on a project that will enable integrating robots into our everyday life. We believe that a "structured environment" (SE) will be one of most important concepts for this project and will also assist robot recognition or movement planning. Now, we focus on a global positioning system (GPS), which is a global SE that gives users or robots their positions whenever and wherever they are outdoors all over the world. GPS will strongly and globally support robot self-positioning. However, GPS has the problem that it cannot be used when the robots are indoors. To solve this problem, we experimentally mounted four pseudolites in our laboratory and developed indoor GPS. The system worked well unless the robot was near the wall, where cycle slip often occurred. To examine the characteristics and reason for cycle slip, we measured the radio-wave environment in the laboratory.

View full abstract

In this paper, we propose a position estimation method for sensor nodes from communication connectivity condition of local communication between the nodes. Actual positions of the nodes are unknown. A host computer, which can collect all of the connectivity conditions, iteratively calculate the estimated positions to be less error of connectivity condition between that of actual positions and that of estimated positions. With our proposed method, spring model, that is giving attracting force or repulsive force among the robots under the difference of their connection conditions between estimated position and actual position. The reasonable positions are estimated by proposed method.

View full abstract

In this research, we develop an automatic driving system using for a compact electric car "i-mobility". The system supports an elderly person and a handicapped person to move somewhere, "i-mobility" avoids obstacles by using environment information and Laser-Range-Finder"LRF" for safety, "i-mobility" provides a service considering situation by using environment information which is embedded in a space.

View full abstract

This paper proposes a simultaneous localization and mapping method in unknown environments for mobile robots. In this research, we used a laser range finder data of the robot for building a map. When the difference between the measured distance and its corresponding map data is large, the robot updates the self-location by using the steady-state genetic algorithm, and updates the map. We propose map building methods based on a topological map. Experimental result shows the availability of the proposed methods.

View full abstract

We propose a robot FOR a pet, not for an owner. Our ultimate goal is to develop a robot that contributes to create and maintain better relation between the pet and the owner while keeping good pet's health condition. In this paper, we propose the robot for the pet and discuss current situation of pet industry. We investigate literatures in the context of animal robot interaction and we discuss a possibility to create a new home robot market. Finally, we show an initial example of ongoing interaction experiment using a tele-operated robot with the dog.

View full abstract

Recently, the practical use of a home robot that can accomplish various tasks in a house is expected. Most of the home robots that have been put to practical use so far perform just a simple task, or are used just as communication tools. We are developing a multifunctional home robot based on the 3M concept (Multitask, Modularization, Minimalization), targeting mainly single company employee type households. A GUI is required to facilitate input of a priori information, e. g. furniture location and desired paths by the user. We propose a GUI which is interactive, intuitive and easy to operate.

View full abstract

We are developing a human-symbiotic robot, called "EMIEW2", designed to support human activity. We set the following three features in the EMIEW2. The first is speed. The robot's moving speed is faster than human walking at 6km/h. The second is the safety. It weighs 13kg the same as a two-year old child. The third is its minimum size for office work. The robot's height is 800mm, which is taller than most office desk. This allows the robot to detect objects on them. To expand on the robot's range of motion, the EMIEW2 has three modes. In "two-wheeled mode", the EMIW2 can move nimbly at 6km/h. In the "four-wheeled mode", the EMIEW2 can move stably. In "step mode", the EMIEW2 can walk up a 3cm step. This paper describes the EMIEW2's mechanism and system constitution.

View full abstract

This paper describes the software architecture of the humanoid robot "EMIEW 2." In order to increase efficiency of software development of robotic and mechatronic systems, which often requires complicated software comprised of many real-time tasks running on multiple processors, the authors have developed a message-driven software component framework. The software of EMIEW 2 is comprised of 28 software components such as motor control, path planning, and localization. The component framework transmits a message containing 1024 bytes of data in 18μs when it is executed on a 500 MHz CPU. It can currently be used with four operating systems and is expected to be applied to other robotic and mechatronic systems.

View full abstract

Nowadays, home-use robots that clean floors automatically are developed. However, the efficiency of these robots is not so good. That is because the best way of cleaning room depends on the room's shape. If the robot knows which room it is cleaning now, it will be possible to make a robot cleaning the room more efficiently. Therefore we propose a method which the robot goes around the room with hug motion, to detect the outline of the room.

View full abstract

Recently, robot cleaner used at home is developing popular. However, because the size is large many of cleaning robots are unsuitable when it cleans the room where furniture is a lot and is narrow. Moreover, these robots are mostly expensive and not easy. Then, this research aims at the development of the small cleaning robot that can easily clean on the corner, chest of drawers, and the desk, etc. in the room etc. in the automatic operation. In this paper we describes about the result of the design, mechanism, and the cleaning experiment.

View full abstract

This research aims at acquiring useful knowledge for robot tasks from many teaching examples by users. We, therefore, develop a system which enables users to easily teach tasks to robots in a virtual world and apply it to acquisition of knowledge for block assembly tasks. Easiness of teaching and a realistic simulation are the keys to a comfortable task teaching in a virtual world. For an easy teaching, the user indicates only blocks and assembly steps by mouse pointing, and the system then automatically calculates the rest of necessary information such as the hand pose for grasping and the robot trajectory. The system also performs a physical simulation for a realistic simulation. Moreover, the interface of the system is implemented using Java for a possible public release in the future. Next, we collect many teaching samples by several users for various block assembly tasks. We then extract useful task knowledge from the samples using machine learning techniques.

View full abstract

This paper presents a method for multi-robot task planning: planning of a mobile robot (parent robot) cooperating with robots which remove obstacles on the parent path (child robots). By using a grid map, this method (1) simulates movement of each robot, (2) estimates total time of travel of the parent robot and (3) optimizes the assignment of each robot's task to minimize the total time. To be robust to unexpected environmental changes in a room, the planning is updated regularly. Our experiment using physical robots showed that the parent robot arrived at the goal cooperating with child robots though an unpredictable obstacle blocked parent's path during its travel.

View full abstract

In this paper, we propose a flexible searching method for the object with a vague information from the people living together with a humanoid robot. And by using the humanoid robot, three dimensional active searching with sight changing could be executed.

View full abstract

Sweeping task means motions that a mobile robot covers a 2-dimention area by its motion. Sweeping algorithms have two categories; one is model-based algorithm and another is reaction-based algorithm. When the robot doesn't have any maps of its workspace, sweeping task must be done by reaction-based algorithm which typically includes some motion primitives. Motion primitives basically consist of sweeping motion and travelling motion. As the sweeping motions, parallel and spiral motions are often used. Following motion and reflection motion for wall are also often used as the travelling motion. This paper investigates a sweeping efficiency using such motion primitives. In simulations, workspaces are modeled by many examples of real rooms.

View full abstract

The paper describes online voice command recognition system using a microphone array. The microphone array attached on ceiling localizes sound source direction in azimuth and elevation, then separates multiple sound sources to recognize voice command in noisy housing environment. These sound localization, separation and recognition functions are implemented as online processing in real world. We define the four indexes to evaluate the performance of the recognition system, and the efficiency in noisy environment or distant sound sources are confirmed from experiments in varied conditions. Finally, an application for mobile robot interface is performed.

View full abstract

The paper describes design optimization method of microphone array using beam forming simulation. Delay and Sum Beam Forming : DSBF is widely used technique to localize and separate sound sources (using microphone array),The performance of DSBF is changes significantly with microphone arrangement, but side lobes are well known problem .We define four indexes to describe directivity of beam forming and evaluation function showing the performance. Then we proposed beam forming simulator with arrangement evaluation function to optimize microphone array design. Finally, 128ch lower side lobes microphone array for mobile robot is automatically determined, and performed over 20dB of S/N ratio.

View full abstract

In this paper, using web database, we build up knowledge database system which both humanoid and human can accumulate information on which humanoid act variably. For application, we construct "face database", and realize behavior for momorised human's face. We will shows how web database is useful for knowledge strage of humanoid.

View full abstract

A versatile system for home service using a mobile manipulator and multi-sensors is being developed by authors, to realize efficient mobile manipulations of various objects in the home by sensor fusion. This paper describes the configuration of the system and mobile manipulation of chairs by utilizing stereo camera and RFID.

View full abstract

We introduce two approachs in Kanagawa. One is "Robo LDK". In a word, it is a approach to think about the usage of the robot by the contest form. ("Contest of living that coexists with robot") The second is support of the proof experiment. ("Robot park project") These are the approachs of verifying needs for the robot by researching user's opinion, and searching for the image of an actually useful robot. I want to institute the necessity of this technique.

View full abstract

This paper describes a passive aerial acrobat robot without electric power. As the robot performs hand-to-hand on its flying trapeze, it simply draws power from the mechanical potential energy of its swings. The robot can transfer continuously throughout a set of the swings that go upward.

View full abstract

In this research, a hemispherical autonomous robot equipped with a large number of LEDs has been developed and the new entertainment system has been created. Authors have developed the robot that a lot of LEDs are mounted on the surface for three years, and have solved some technical problems for these robots. Characteristics and technical problems of each robot are discussed in this thesis. In the first year, the fundamental concept of robot was designed, and the prototype of the robot was created. We worked on the development of the luminescence control system of full color LEDs and the intercommunication system using wireless LAN network in the second year. We have examined the motion path correction method that uses the Bezier curve and the system by which two or more robots obtain their position information from the image processing system. Experimental results are presented to illustrate the validity of the present robot for the new entertainment system.

View full abstract

This study aims at developing a violin playing robot that is able to produce expressive sounds. This will lead us to understanding human Kansei. In this paper, we describe a volume feedback system for an anthropomorphic violin playing robot with 7 D.O.F., which we have been developing. This system imitates the human procedure to improve their playing musical instruments, which is affected by human Kansei. We analyzed the relationship between bow speed and volume of the sound produced by the robot. Then, we constructed an algorithm for estimating bow speed needed to produce indicated volume sounds. We confirmed that the robot was able to produce sounds with indicated volume using the algorithm.

View full abstract

In this project, a robot, that models "SUZUHURI(laughing man)" who acts in "WARAI religious festival" held at Niu Shinto shrine in Hidaka-gawa town, was developed. The developed robot was called WARAI Robot. Head and arms of Bunraku puppet was applied as those of the developed robot, because Hidaka-gawa town is famous for Bunraku. WARAI Robot was demonstrated in several events. This project attracted many media, and it was reported in newspapers and TV shows. As a result, this project contributed to activation of Hidaka-gawa town.

View full abstract

Magic is a performance unique to humans that makes use of hand's skills to its maximum extent. It needs enormous efforts to acquire the skills of magic for a skillful man. In this study, a robot hand that can handle a small object in the palm and perform coin magic making full use of its palm and fingers. It is said that the coin magic is one of the most difficult kinds of magic. Therefore, we expect that a robot that can do coin magic may be able to do any magic successfully. In this paper, a mechanism of the robot hand is studied and the control methods of handling coins for the magic is discussed. The robot is assumed to use a Japanese 500 yen coin during the performance and it is assumed that the robot controls the hand by using only the information from the palm sensors. It is the problem how to design the skilful hand, how to construct a control algorithm, and what sensors should be used to sense the state variables of the coin.

View full abstract

Recently, it becomes very important to make the environment of handicapped people in the daily life more comfortable and there are many rising movements and activities to fulfill the subject. As a result, the occasion and the field where handicapped people can exercise their ability are increasing. But it is very difficult for them to play such as the wind instruments freely, because such performance needs very skillful and delicate movements of their body parts. In this research, we assume that a professional sax player became unable to use his lips to make good embouchure because of the lack of the lip muscle force and unable to play the saxophone properly. The research object is that we should develop an easily controllable device to make artificial embouchure and the device should enable the player to play the saxophone as before.

View full abstract

This is a study on a singer robot using artificial larynx. The robot can generate Japanese 5 vowels [a], [i], [u], [e], and [o], in the present. It is very difficult to generate clear vowels because the frequencies of formants are not easy to be controlled. Formant frequencies can be manipulated by changing the cross-sectional area of the vocal tract. Singer Robot is expected to produce vowels with correct formants in order to sing in distinct pronunciations. We propose a new adjustment system of vowel formant. This system can search an optimal cross-sectional area distribution of vocal tract for the given magnitude of the formant frequency change. And, the system introduced in this paper analyzes formants in real-time and feeds back them to Singer Robot.

View full abstract

Recently researches of biped running robots are increasing. These researches deal with low-velocity running such as jogging. As the running velocity becomes higher, the ground reaction force including impact force at touch down becomes larger than that of low-velocity running. The large reaction force makes the robot unstable. Therefore, in high-speed running, the control of the ground reaction force is very important. In this study, we modeled the ground as a dynamical model of two springs and a dashpot. The model can make the reaction force curve during the landing period just the same as that of a human athlete. The trajectories of the joint angles are derived so that the mass center of the whole body might change like a human athlete during the landing period.

View full abstract

This paper describes a technique of a beat tracking that recognizes temporal positions of A cappella song. The technique is divided into four procedure; acquisition of data, FFT analysis, temporal-positions recognition and rhythm recognition. The temporal-positions recognition utilizes that frequency spectra change suddenly at temporal positions. The rhythm recognition is based on the eight note, and finds out the tempo of a song when the error margin at a theoretical and an obtained temporal position is small. It was found from the experimental result that the technique is able to recognize the rhythm of the voice synthesis song by means of performing FFT analysis to 256 data every 200 or 300 times.

View full abstract

In RoboCup middle size league, the official rules have been changing year by year. In 2008, goal's colors (blue, yellow) changes no color goal such as human soccer goal. Furthermore, the ball changes orange ball with FIFA official balls. Therefore, in the middle size league, it has being very difficult to recognize the objects. In this research, we developed ball recognition algorithms, because most teams are using a method of orange color extraction for recognizing ball now. This paper shows experimental results of two ball recognition algorithms. These algorithms are Hough transformation method, and ball estimation method by green inversion.

View full abstract

In this paper we present a method of self localization in the RoboCup middle size league based on the field lines. This method estimates position of the robot by using Monte Carlo Localization (MCL). MCL is widely used for mobile robots localization in a known environment due to its good real-time performance and robustness. MCL estimates state of a robot based on sensory data. In our algorithm, the sensory data are distances to field lines, wheels odometry measurements and direction sensor data. It is prone to estimation error in RoboCup games because robots move around fast in the field and crash each other. So we also study recovery methods for estimation error.

View full abstract

Recently, many kinds of robots such as industrial and service robots has being developed. The International Organization of Standardization (ISO) provides the principle of machine design in general, by ISO12100[1]. The engineers should reduce the risk based on principle of safe design, so that they could satisfy the demands from society. That is to say, a service robot can be accepted and used in society, if designers of the robot can satisfy the principle of safe design. Regarding to the robots which are used in RoboCup Soccer it have the problem that a human may crash into the robot. Therefore, as one of principles of safe design (safe in crashing robot with human), we should consider the safety problem of the robot that is used in RoboCup Soccer as well. In this paper, we introduce the process that is how to reduce the risk.

View full abstract

RoboCup is a competition where autonomous mobile robots play soccer. It was initiated to encourage development of intelligent robots. In the RoboCup, a role of goalie (a goalkeeper robot) is to prevent score by the shot from an opponent robot. For the goalie it is demanded to be highly precise, and to move quickly to the ball course to minimize shoot course. In this paper we present algorithm of goalie self-positioning, prediction of the ball course, calculation of the ball height in a loop shot, and present control algorithm for the goalie movement to a specified point in front of the goal.

View full abstract

This paper presents a vision system for an outdoor type robot in the RoboCup Middle-Size League (MSL). We use the image processing library called "Intel Open Source Computer Vision Library" (OpenCV) to develop this vision system. In this method, first, the system gets a coordinate of an object by template matching or color information. Second, a template is generated in real time to recognize an object from an image of a previous flame. This template is used for object recognition from next flame. As a result, it will be able to repress to degree of dependence of color information. We demonstrate effectiveness by testing this vision system in some lighting conditions which assume outdoor game.

View full abstract

In this paper, we introduce autonomous humanoid robot system "CIT Brains and Hajime Robot" for RoboCup Humanoid League. To play soccer game autonomously, humanoid robot should have many functions such as recognizing situation, selecting behavior and controlling motion. We develop robot mechanism, controller for walking, image processing, algorithm of selecting behavior, communication, and so on. We implement such wide area technologies to our system. We describe our system and show findings though development.

View full abstract

We are making research on a kicking mechanism that can kick a ball like a professional soccer player. We measured human kicking motion by using high speed camera as well as electromyograph sensors and analyzed the motion of knee and ankle. Based on the modeling as a two-link mechanism, we decided the design parameters for the control of the kicking mechanism, and manufactured a first prototype kicking machine. Next, we simulated human kicking motion using MATLAB/Simulink, then we made some real kicking experiments using the same motion algorithm. Currently, the maximum kicking speed is set below the desired value so that the kicking impact force may not damage to the mechanism. We show these design considerations of this first prototype of kicking machine in this report.

View full abstract

We started to participate in RoboCup humanoid league. Based on the previous experience, more lightweight robots are needed for this application. Then we adopted a commercial legged robot "e-nuvo" by ZMP Inc., and added two arms as well as a camera, a board PC and batteries. Total weight of the new humanoid robot is 1.8kg, which is only 0.5kg bigger than the weight of the original e-nuvo. Two arms are driven by the same processor with other leg angles and incorporated in the same architecture, so that the arms are controlled in the same software system. We adopted OpenCV on Windows for the image processing library, and implemented image recognition modules to detect ball and goal positions. We are now developing the total software system that is sufficient for the RoboCup competition.

View full abstract

Kawagoe festival is the traditional Japanese festival that is held in Kawagoe, Saitama, where multiple "Dashi's", or floats, move around the city center with music and dance on each of them. We have decided to simulate this festival by using small simple robot kits and to hold a robot contest for high school students. This event includes preparatory seminars and contest itself, and students plan their ideas in the technical part and artistic part of robot design. The purpose of this contest is to let high students know the traditional Japanese culture as well as to get technical interests on robots. This paper reports the philosophy of holding a robot contest for high school students, technical rules and regulations, as well as some considerations of this event.

View full abstract

The paper describes development of a variable kick mechanism for RoboCup soccer agent. Most robots entering to the middle size league of RoboCup equip a kick mechanism, and many of them mount loop kick device in recent year. The loop kick is not suitable for a short pass though has the ability to score goals. Therefore, various cooperative plays between robots are difficult only for using loop kick. To realize various cooperative strategies, we propose variable kick mechanism. The mechanism changes the contact point with the ball by changing the position of the tiptoe. Then, the trajectory of the ball is changed. Experiments confirmed that both the loop kick and the straight pass were possible by proposed mechanism.

View full abstract

We introduce the competition on the plan of robotics and mechatronics product with evaluating its engineering function and interface design. Fifteen teams of six universities joined the competition on the subject, "Supporting system of health, care and communication of the elderly." Many of the plans presented the ideas with the balance on both function and design, showing the results of the collaboration by the students of engineering course and of product design course.

View full abstract

Both self-learning architecture (embedded structure) and explicit/implicit teaching from other agents (environmental design issue) are necessary not only for one behavior learning but more seriously for life-time behavior learning. This paper presents a method for a robot to understand unfamiliar behaviors shown by others through the collaboration between behavior acquisition and recognition of observed behaviors, where the state value has an important role not simply for behavior acquisition (reinforcement learning) but also for behavior recognition (observation). That is, the state value updates can be accelerated by observation without real trials and errors while the learned values enrich the recognition system since it is based on estimation of the state value of the observed behavior. The validity of the proposed method is shown by applying it to a dynamic environment where two robots play soccer.

View full abstract

The Variably Configured Segmented Wheel (VCSW) seems to be suitable for traveling on snowy road. The tip of vane is effective to prevent slippage on a snow road and the motion of vanes is effective to remove covered snow. The amount and direction of vane extrusion should be controlled according to the road condition. Three motors for each wheel should be controlled. An on-board controller with a micro computer having wireless telecommunication device is developed for multi axis control. This controller is applied to the test vehicle successfully. But the power control should be improved.

View full abstract

My laboratory attempts to create a new rescue system with heavy machinery. This system is supposed to be made by replacing power shovel's bucket with dual-arm manipulator. This system has two types of operational method. One type is remote control type called "Master-Slave type". The other type is called "Power-Assist type". For the first step to create the system,I make a study on Power-Assist type. In Power-Assist type,operator operates the system on dual-arm manipulator. If operator sway in operating,the system has to follow him. So,in this study,I control the motion of system by impedance control.

View full abstract

Our research group has developed a tracked mobile robot "kenaf' for high-speed exploration of the confined space such as the underground cities. The kenaf can move about 3 km/h and can get over about 40 cm step. However, it is difficult to pass through the steps and the narrow spaces quickly by remote control. The authors propose an autonomous speed adjustment system based on measurement of environmental shapes. The system limits the maximum speed of the robot when the obstacles are detected within the kinematics-based ellipsoidal areas. The system also limits the speed when the environmental shape ahead is the steps or the slope. The distances of the obstacles were measured by using the laser range sensor. The environmental shape ahead was obtained by using the range image sensor. The validity of the system was experimentally evaluated in a narrow space and a step.

View full abstract

The authors aim at development of a rescue crawler robot to search for victims in rubble environment. A tracked robot with flippers "Sinbad" was built for the purpose. The robot has 6 degrees of freedom (DOF) for getting over the rubble. However, it is too hard for an operator at the remote place to control these multi degrees of freedom. In this paper, the authors will explain our approach of the remote control support for the crawler robot with flipper arms. Especially, the authors will explain about our developed distributed touch sensors for the crawler robot, and propose a method of rollover avoidance based on the distributed touch sensor.

View full abstract

When we operate rescue robots in desarster area, the most effective tool is visual infomation obtained by video cameras on the robots. However, the network transmission capacity in a traditional single channel Ad-hoc network is limited to 4-6Mbps in general. This amount of capacity is not enough because rescue robot mounts 2 or 3 cameras to gather enviromental infomation. This paper describes the development of robots with multiple channels Ad-hoc networking system to extend the scale of rescue operation.

View full abstract

To enhance the communication distance of WLAN, multi-hop network is expected as one effective method, which means that the robot grounds the devices which relay the communication or one robot relay the communication from another robot. In this paper, we describe experimental result of our long distance navigation of rescue robots with multi-hop network.

View full abstract

This research is concerned with development of a rescue robot that search people buried in debris when an earthquake happens. "Foot Groping" is a key operation to walk continuously on rubble, which is shown in estimating ground. The purpose of this research is to clearly establish a relation between actuations and sensors by using an experimental robot RoQ with foot groping skill extracted by human. In this paper, We do the new suggestion of a legged robot which considered Foot Groping.

View full abstract