Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

In this paper, we propose a spatial measurement method based on object geometry and velocity information using past measurement information from 3D-LiDAR. If geometric information is obtained from 3D-LiDAR, it can be applied to global localization, object identification, etc., and robots will be able to run autonomously with greater stability. However, 3D-LiDAR cannot always measure point densities high enough to acquire sufficient geometric information. Conventional methods have been used to control point density by interpolating areas of low point density or by varying the rotational scanning speed of 3D-LiDAR. However, these methods have problems such as the shape of the measured object differing from the real object and the loss of real-time measurement. In this paper, therefore, a remaining time is set for each point based on the shape information of the object, and each point is allowed to remain until the remaining time is reached, depending on the robot’s movement. The remaining time is set so that the point density is preferentially increased in areas where shape information is estimated to be abundant and in distant areas where the point density tends to decrease due to the LiDAR mechanism. For dynamic point clouds, velocity information is added, and the point cloud is moved to the estimated position the next time and superimposed on the latest dynamic point cloud while considering the distance between points, thereby improving point density while suppressing shape collapse. We demonstrate the usefulness of this method through experiments on actual equipment.

DOI： 10.1299/transjsme.23-00034

CiNii Research

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII55687.2023.10039292

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2023.html#KawaiK23

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII55687.2023.10039123

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2023.html#HirakawaK23

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1177/17298806221129842

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII52469.2022.9708864

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2022.html#KawaiYHSMK22

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII52469.2022.9708741

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2022.html#HirakawaK22

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1186/s40648-021-00213-5

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1186/s40648-021-00196-3

Web of Science

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

This paper presents a real-time self-attitude estimation method which utilizes the clues to the direction of the gravity hidden in images and structures of the environments. In the proposed method, the angular velocity is integrated using a gyroscope, a camera-based method estimates the gravity direction, and a LiDAR-based one also estimates the gravity direction, respectively. These estimations are integrated using the EKF (extended Kalman filter). The camera-based gravity direction estimation uses a DNN (deep neural network) which learns the regularity between the gravity direction and the landscape information. By learning the regularity, the proposed DNN can infer the gravity direction from only a single shot image. The DNN outputs the mean and variance to express uncertainty of the inference. The LiDAR-based gravity direction estimation extracts vertical planes from the surrounding environment measured by the LiDAR, and outputs the gravity direction based on their normals. By using both the camera and the LiDAR, more robust and accurate estimation can be achieved. To show the DNN can estimate the direction of gravity with uncertainty expression, static validations on test datasets are performed. Dynamic validations are also performed to show the proposed EKF-based method can estimate the attitude in real time. These validations are performed in both simulator and real world to compare the proposed method with conventional methods.

DOI： 10.1299/transjsme.21-00098

DOI： 10.1299/transjsme.22-00325_references_DOI_8RBkKBtPQbhOJmelyyXMmdXuGnf

CiNii Research

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IEEECONF49454.2021.9382642

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2021.html#OzakiK21

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1109/LRA.2021.3060442

Web of Science

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

In this paper, we propose a pedestrian trajectory prediction method for autonomous mobile robots. In many cases, there are many pedestrians in the environment in which the autonomous mobile robot runs. In such an environment, the robot needs to run safely for pedestrians. In order to avoid collisions with pedestrians and drive safely, it is important to predict future movements of pedestrians. In the conventional prediction method, the trajectory of a future pedestrian is often predicted from the position of a pedestrian in the past. However, in such cases, it is difficult to predict the movement to avoid obstacles such as walls and pillars around the pedestrian. In this study, point cloud data acquired by LiDAR is used to predict the behavior of pedestrians avoiding surrounding static obstacles. Based on the point cloud data, distances between the pedestrians and the static obstacles is calculated at each time. Then, input it into the network together with the transition of the pedestrian’s position to predict the future pedestrian’s position. In addition, we use attention mechanisms to model interactions between pedestrians. This makes predictions that consider not only static obstacles but also the effects of other pedestrians. The usefulness of this study is shown by performing accuracy evaluation using the dataset created in the simulation environment and the publicly available dataset.

DOI： 10.1299/transjsme.21-00125

CiNii Research

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII46433.2020.9026219

Web of Science

Scopus

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2020.html#NagaiKK20

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2020.p1112

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2020.p1104

Web of Science

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

When a mobile robot moves in an environment where there are moving obstacles such as pedestrians and other robots, the robot is required to avoid collisions with these obstacles. However, trying to avoid all the predicted positions on the trajectory of the moving obstacle would overly restrict the robot’s movement. In order to solve this problem, we should be able to generate trajectories in which the robot and moving obstacles do not approach at the same time in the future. However, an approach such as optimizing the robot’s trajectory for the predicted trajectory of a moving obstacle would result in a very large number of states to be considered. This paper proposes a method to generate a set of candidate trajectories based on the robot’s motion model and determine the trajectories considering the collision probability with the predicted trajectories of moving obstacles at each time. In the proposed method, the action of the robot is determined by the following process. First, candidate trajectories of the robot based on its motion model are generated. Next, at each time step of each candidate trajectory, the collision probability with each obstacle is calculated. The same calculation is made for all obstacles and the collision probability with one or more obstacles is calculated. If the maximum probability of a candidate trajectory is greater than the threshold, the trajectory is considered dangerous and is removed from the candidate list. Depending on the distance from the target, the robot selects a trajectory from among the remaining candidate trajectories and determines the next action. In a simulation environment with 30 moving obstacles, the performance is compared with other methods and shown to be comparable or better. In addition, experiments using a real robot were conducted to show the usefulness of the proposed method.

DOI： 10.1299/transjsme.20-00102

CiNii Research

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII46433.2020.9026222

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2020.html#OzakiK20

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII.2019.8700333

Web of Science

Scopus

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2019.html#SadakuniKOK19

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

In the conventional intersection recognition method, shape information is used. In order to recognize an intersection composed of roads divided by grass and asphalt, it is necessary to distinguish them. There is almost no difference in shape between grass and asphalt. For this reason, it is difficult to distinguish them by the method using shape information. Therefore, we use reflection intensity by LiDAR to distinguish grass and asphalt. In this paper, we compare the reflection intensity of the ground surface at points where distances from LiDAR are equal. The appropriate threshold for distinguishing materials of the ground is dynamically calculated for each distance using Discriminant Analysis Method. We accumulate the drivable region using the dynamically calculated threshold value in chronological order taking the relative movement amount into account. We add probabilistic processing to the drivable region accumulated in chronological order to extract more stable drivable region. Finally, an intersection is detected by combining the drivable region extracted using the reflection intensity and the drivable region extracted using the shape information. Intersection detection method uses Toe-Finding Algorithm. In order to show the usefulness of our method, we compare our method with two method, Partition Line Method and the proposed method unaccumulated the drivable region.

DOI： 10.1299/transjsme.19-00064

DOI： 10.20965/jrm.2023.p0347_references_DOI_NeJoOqy0u2J2Cx2KYR2AwOoBOUa

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

This paper presents a real-time 6DoF localization method which corrects accumulative error by estimating relative poses to building walls for mobile robots in urban areas. This method exploits a fact that most of all artificial walls are built vertically. It estimates poses by not only an inertial sensor but also real-time SLAM and observations of normals from point-cloud of artificial walls for estimating absolute poses in the gravity coordinate system. Those three types of poses which are estimated by each way are combined by extended Kalman filter. To evaluate the proposed method, outdoor experiments with an actual robot were performed. They show the method keeps correcting accumulative error while the robot moves. The other experiments show how the method suppresses influence of non-vertical planes.

DOI： 10.1299/transjsme.19-00065

DOI： 10.1299/transjsme.21-00098_references_DOI_SbXe3e25oJTZ4k07RBi2PkHF7qs

CiNii Research

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2018.p0552

Web of Science

researchmap

Other Link： https://dblp.org/db/journals/jrm/jrm30.html#HosodaSMYSOKKOK18

Publishing type：Research paper (international conference proceedings)   Publisher：International Federation of Automatic Control  

DOI： 10.1016/j.ifacol.2018.11.542

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Publishing type：Research paper (international conference proceedings)   Publisher：International Federation of Automatic Control  

DOI： 10.1016/j.ifacol.2018.11.579

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Publishing type：Research paper (international conference proceedings)   Publisher：International Federation of Automatic Control  

DOI： 10.1016/j.ifacol.2018.11.580

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2017.p0639

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2015.p0401

Web of Science

CiNii Research

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：the Society of Instrument and Control Engineers  

In this paper, we propose a mobile robot localization system in frequent GPS-denied situations. We utilize multiple observations that are obtained from sequential appearance-based place recognition and GPS. Using GPS observations has still some challenging problems such as multipath or signal lost under environments where there are tall buildings nearby. The appearance-based place recognition that is combined with positional information has capability to overcome the issue. We apply both of observations derived from GPS and appearance-based place recognition to a mobile robot localization for the sake of achieving robust localization. Moreover sequential appearance-based place recognition makes it possible to recognize their own position even when we navigate a robot at night. Our system uses not only multiple observations but also dead reckoning with gyrodometry model. Our experiments are performed over aggregate 5300m trajectory approximately that contain a 1600m outdoor route in different seasons and at different times, and a 500 m short-range route to verify its validity.

DOI： 10.9746/sicetr.51.57

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Society of Instrument and Control Engineers  

In this paper, we describe a versatile navigation system for autonomous mobile robot that we have developed. As a development background, there is an increasing demand for autonomous mobile robot to play an active part in the human living environment in recent years. That is indicated to need a more versatile to the navigation system. In here, the versatile system that it can be used for indoor or outdoor without limiting the operating environment. To realize this, we propose a navigation system using an electronic map which is easy to access. To perform the localization by comparing the information of shape about the surrounding environment acquired from LIDAR data with electronic map data around the robot's position. In order not to approach into a local hazardous area where are not registered on the electronic map, robots have to detect and avoid the area on the fly. To detect the hazardous area, traversability analysis should be performed in real-time with a 3D point cloud acquiring form the LiDAR data. The collision avoidance maneuver may be carried out by a real-time path planning based on the dynamics of the robot and the results of the traversability analysis. Several experiments along with the route which contains both indoor and outdoor environments in the distance of more than 1,300m travel are carried out to examine the system's versatility.

DOI： 10.9746/sicetr.51.545

CiNii Research

J-GLOBAL

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I026693976

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2014.p0185

Web of Science

CiNii Research

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICRA.2014.6907431

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2014.html#SaitoKK14

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII.2013.6776624

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ICRA.2013.6630776

Web of Science

Scopus

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Robotics Society of Japan  

In this paper, the autonomous navigation system which traverses in complex urban area without prior environmental information. If very large scale information (e.g., scanned three dimensional points cloud) on the route were known priorly, autonomous navigation would be easy. However, if robots had to move long range in the area where is out of control, quite few information could be given previously. In ''Tsukuba Challenge'', it is allowed that capturing any kind of information on the route beforehand and using it on the challenge day. Since autonomous navigation system is valid in realistic situation, we try not to use such prior information for the navigation system. Therefore, we propose an autonomous navigation system which only uses sensory data valid in real situation, e.g., positions from GPS, attitude from IMU, visual images and LRF data. Especially for visual images and LRF data, we only use them for on-line processing, or we handle them as one of the general information. Road boundary estimation and speed control were successfully executed in the presence of many pedestrians, other robots, and variance of road shape. The effectiveness of the proposed system is proved through the final result of the ''Tsukuba Challenge 2010''.

DOI： 10.7210/jrsj.30.907

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Japanese Society for Planetary Sciences  

DOI： 10.14909/yuseijin.21.3_206

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I024008715

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2012.p0016

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/SII.2012.6426958

Web of Science

researchmap

Language：English  

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I11081012

Language：English  

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I11080986

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE/SICE  

DOI： 10.1109/SII.2011.6147516

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/ROBIO.2011.6181425

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/AIM.2011.6027041

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/AIM.2011.6027045

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE/SICE  

DOI： 10.1109/SII.2011.6147636

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/ROBIO.2011.6181424

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2010.p0726

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：VDE Verlag  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：VDE Verlag  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：VDE Verlag  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2010.p0140

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2010.5651185

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2010.html#SakaiMK10

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/ROBIO.2010.5723383

Scopus

researchmap

Other Link： https://dblp.org/db/conf/robio/robio2010.html#MitsuhashiSK10

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/ROBIO.2010.5723457

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ISCIT.2010.5665142

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ISCIT.2010.5665126

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IFAC Secretariat  

DOI： 10.3182/20100913-3-US-2015.00066

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IFAC Secretariat  

DOI： 10.3182/20100913-3-US-2015.00064

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IFAC Secretariat  

DOI： 10.3182/20100913-3-US-2015.00063

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1163/016918609X12619993300746

Web of Science

researchmap

Other Link： https://dblp.org/db/journals/ar/ar24.html#SaitohSK10

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1163/016918610X501273

Web of Science

researchmap

Other Link： https://dblp.org/db/journals/ar/ar24.html#KurodaST10

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2010.5651208

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2010.html#SaitohK10

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/RIISS.2009.4937903

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/riiss/riiss2009.html#SaitohSK09

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2009.5354516

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2009.html#TamuraSIK09

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2009.5354677

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2009.html#SakaiTK09

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/RIISS.2009.4937905

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/riiss/riiss2009.html#KurodaS09

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1017/S0263574706003171

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1002/rob.20118

Web of Science

researchmap

Publishing type：Research paper (scientific journal)   Publisher：Elsevier BV  

DOI： 10.1016/j.asr.2005.05.097

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/AMC.2006.1631669

Web of Science

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

In future planetary exploration missions, planetary robots are required to traverse over very rough terrain. On the lunar surface, there are covered with soils which are named regolith. Regolith is easy to slide. Any stack is easy to occur to the rovers which are running on the regolith. The mechanism, which occurs stacks, is not yet clear. Therefore, this paper investigates the kinematics behavior of lunar rovers with tire-soil traction mechanism. This traction mechanism shows two important parameters for running wheel on regolith. One is the effect of hardening. And the other is the effect of shearing. New wheel forms are designed by considering these effects. Some experiments using the proposed wheels are performed to observe the performance of traversability. Analysis of experimental results shows that to combine these two effects is effective.

DOI： 10.1299/kikaic.72.3842

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I8629267

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2006.281956

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2006.html#IizukaSKK06

Language：English   Publishing type：Research paper (scientific journal)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ROBOT.2005.1570542

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2005.html#ShimodaKI05

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2005.1545177

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2005.html#KurodaTSK05

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ROBOT.2004.1308133

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2004-2.html#KurodaTSK04

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ROBOT.2004.1308132

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2004-2.html#KurodaKTK04

Language：English   Publishing type：Research paper (scientific journal)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2003.1249256

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2003.html#KurodaKM03

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1163/156855302317413718

Web of Science

researchmap

Other Link： https://dblp.org/db/journals/ar/ar16.html#KumagaiUKW02

Publishing type：Research paper (scientific journal)  

DOI： 10.1016/S0273-1177(02)00491-X

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ROBOT.2001.932805

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2001.html#KuniiSKK01

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.2001.973343

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros2001.html#KuniiTKK01

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/ROBOT.2001.933195

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/icra/icra2001.html#KubotaKKY01

Language：Japanese   Publishing type：Research paper (scientific journal)  

CiNii Research

researchmap

Language：Japanese   Publisher：The Institute of Electrical Engineers of Japan  

DOI： 10.1541/ieejjournal.120.754

CiNii Research

researchmap

Other Link： https://jlc.jst.go.jp/DN/JALC/00070220735?from=CiNii

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.20965/jrm.2000.p0091

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.1163/156855300741861

Web of Science

researchmap

Other Link： https://dblp.org/db/journals/ar/ar14.html#KubotaKKN00

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： 10.1109/IROS.1998.724682

Web of Science

researchmap

Other Link： https://dblp.org/db/conf/iros/iros1998.html#KubotaNKASI98

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

DOI： 10.1109/ROBOT.1998.680616

researchmap

Other Link： https://dblp.org/db/conf/icra/icra1998-4.html#HosokawaTFKAKE98

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

This paper proposes and demonstrates a cable tracking system based on a visual feedback navigation as an application for the environmental survey using AUVs (Autonomous Underwater Vehicles) . The proposed cable tracking system consists of three levels of controllers : (1)The higher level controller decides the AUV's steering mode which includes a mode to search the cable in case that the vehicle loses it. (2)The middle level controller generates target values to realize the steering determined in (1)by fusing vision data and other sensing data. (3)The lower level controller generates control commands for the actuators according to the target values. This Paper also addresses the vision processing techniques including Hough transformation and the transformation from two dimensions to three dimensions which are necessary for the steering decision in(1)and the target generation in(2) . The proposed system is implemented on an actual AUV named "Twm Burger 2 ". The performance of the system was examined though trials carried out at the Lake Biwa on October 1996. The vehicle navigated for about fifteen minutes without losing the cable although the cable was sporadically veiled in waterweed and transparency of water was not good. It is shown that the vehicle implemented the proposed cable tracking system can be a practical and general platform for environmental survey in the predetermined area.

DOI： 10.2534/jjasnaoe1968.1997.213

CiNii Research

researchmap

Other Link： http://dl.ndl.go.jp/info:ndljp/pid/10788962

Language：English   Publishing type：Research paper (international conference proceedings)  

Web of Science

researchmap

Language：Japanese   Publisher：日本造船学会  

収集根拠 : NII-ELS
資料形態 : テキストデータ
コレクション : 国立国会図書館デジタルコレクション > 電子書籍・電子雑誌 > 学術機関 > 学協会

CiNii Research

researchmap

Other Link： https://dl.ndl.go.jp/pid/10788928

Language：Japanese   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

This paper introduces a simulator system essential to the research of underwater robots which are required to have highly reliable and efficient ability of information processing. The universe of the simulation consists of independent worlds which include specific agents such as robots and researchers. The features of the proposed system are as follows : 1) The whole system is operated in real time on the computers distributed in the network including the actual underwater robots. 2) Each agent program in the simulation world is connected to the programs called "managers" which handle the interactions among the agents. 3) Underwater environment is modeled as well as the robot itself. 4) Multi-user simulation is available, so that researchers in remote places can share the same simulation world. With this simulator, for example, the software loaded into a robot can be examined by giving sensor readings in a virtual environment to the actual robot swimming in a testing pool. Since the system continuously provides researchers a virtual underwater environment, they can operate their robots as if they were in the ocean.

DOI： 10.2534/jjasnaoe1968.1995.178_667

CiNii Research

researchmap

Other Link： http://dl.ndl.go.jp/info:ndljp/pid/10788684

Language：Japanese   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

Multiple vehicle operation (e. g. cooperative object search) using autonomous underwater vehicles is an expected paradigm for deep ocean survey and exploration. In this paper, the system which consists of homogeneous vehicles is introduced to make a robust multiple vehicle system. The vehicle control architecture implemented in each member of the vehicles includes two decision making modules : one refers the individual environmental information gotten by its own sensors, and the other shares information received by other vehicles. By separating the decision maker into these two modules, the mission operator can easily write programs for each vehicle's behavior to generate an appropriate total vehicles' behavior. The performance of the vehicle's controlled by the proposed system is demonstrated on a multi-vehicle simulator, and the total behavior shows that the vehicles generate a formation which has not been explicitly given.

DOI： 10.2534/jjasnaoe1968.1995.447

CiNii Research

researchmap

Other Link： http://dl.ndl.go.jp/info:ndljp/pid/10788575

CiNii Research

researchmap

Language：Japanese   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

Software development for intelligent underwater robots is one of the most emerging technologies to support human activities underwater. A versatile test bed named "The Twin-Burger" was constructed in 1992 at the Institute of Industrial Science, the University of Tokyo for software development and verification in realistic environment. The Twin-Burger is a fully autonomous underwater robot which measures 1.54 m long and weighs approximately 120 kg in air. This paper describes the design philosophy of the robot and its hardware and software systems in detail. The Distributed Vehicle Management Architecture (DVMA) is proposed as a software architecture for the Twin-Burger to provide a mechanism to generate a sequence of behaviors for a commanded mission while handling the hardware interfaces in real time. Mission execution experiments were carried out to verify the architecture's effectiveness. It is shown that the Twin-Burger behaves appropriately according to the mission and environmental conditions by the DVMA.

DOI： 10.2534/jjasnaoe1968.1993.174_903

CiNii Research

researchmap

Other Link： https://jlc.jst.go.jp/DN/JALC/00366986932?from=CiNii

Language：Japanese   Publisher：東京 : 日本機械学会 ; 1979-2010  

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I3777764

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Publ by Univ of Tokyo Press  

Scopus

researchmap

CiNii Research

researchmap

Language：Japanese   Publisher：The Japan Society of Mechanical Engineers  

The performance of a damping-controllable dynamic damper was investigated with an adaptive neural net control system. The damping controllability of the dynamic damper was accomplished by introducing the Electro-Rheological (ER) fluid, which is known as one of the highly functional materials and whose apparent viscosity can be varied in accordance with applied electric field strength in a wide range as well as in a short time. As a control system for the present dynamic damper, an adaptive neural net control system was developed ; it was composed of a forward model network and a controller network. With the proposed neural net control system, the damping-controllable dynamic damper can be tuned by itself even if response of a structure shows nonlinearity to input disturbances or varies its own dynamic characteristic. Moreover, the vibration amplitude corresponding to several natural frequencies can be reduced with only a single dynamic damper. Numerical and experimental investigations were performed, and the results shows the superiority of the present adaptive vibration control system.

DOI： 10.1299/kikaic.58.1748

DOI： 10.1299/transjsme.18-00062_references_DOI_9yqlHb5z1J9X510XSFIkVZwTuZO

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

In the previous study, the Self-Organizing Nerural-Net-Controller System (SONCS) was developed as an adaptive control system. In order to utilize this system more efficiently, it is important to investigate modelling ability and learning flexibility of the forward model networks. Under the everchanging environmental condition, the forward model should be able to express the complex dynamics of the controlled object and should be easily modified. In this study, the forward model network is improved by adding recurrent connections from the hidden layer to the input layer instead of recurrent connections in the input layer of the previous one. Characteristics of both the forward model networks are investigated using a simple nonlinear system as the modelled object. It is concluded that the frequency range the forward model can cover is spread by this improvement of the recurrent connections.<BR>A small test-bed, which has been built to be used to test the control system in the real environment, is introduced.The new version of SONCS is tested using this test-bed through free-swimming tank tests. It is shown that the SONCS can appropriately generate and adjust the controller which is intended to let the vehicle swim at the desired depth.

DOI： 10.2534/jjasnaoe1968.1991.477

DOI： 10.2534/jjasnaoe1968.1990.168_275

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I3703479

Language：Japanese   Publishing type：Research paper (scientific journal)   Publisher：The Japan Society of Mechanical Engineers  

Electro-Rheological fluid (ERF) has notable characteristics of viscosity which could be applicable for various industrial machines. Viscosity of ERF increases in accordance with the applied electric field. Without an electric field, ERF has the viscosity characteristics of a Newtonian liquid. Variation of viscosity is reversible, and the required time for its variation is said to be as short as 0.001 sec. Numerous applications of ERF in the engineering field have been attempted since the characteristics of ERF were discovered by W. M. Winslow. However, no successful application has ever been reported. This paper deals with the successful application of ERF to a controllable shock absorber system for automobiles. The controllable range of damping force is examined experimentally for various piston speeds. In addition, numerical experiments are performed; in which damping force can be controlled in response to road conditions by using the shock absorber proposed in this paper. Pitching and rolling motions of the automobile are found to be decreased remarkably.

DOI： 10.1299/kikaic.56.928

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I3661488

Language：English  

CiNii Research

researchmap

Language：Japanese  

researchmap

Language：Japanese  

CiNii Research

researchmap

Other Link： http://hdl.handle.net/10291/13182

Language：Japanese  

CiNii Research

researchmap

Language：Japanese   Publisher：The Japan Society of Naval Architects and Ocean Engineers  

DOI： 10.14856/technom.884.0_240

CiNii Research

researchmap

Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I7291382

Language：Japanese   Publisher：The Robotics Society of Japan  

DOI： 10.7210/jrsj.21.477

CiNii Research

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electrical Engineers of Japan  

DOI： 10.1541/ieejjournal.120.754

CiNii Research

researchmap

Other Link： https://jlc.jst.go.jp/DN/JALC/00070220735?from=CiNii

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

DOI： 10.1163/156855399X00153

researchmap

Other Link： https://dblp.org/db/journals/ar/ar13.html#FujiiK98

Language：Japanese   Publisher：公益社団法人 日本船舶海洋工学会  

DOI： 10.14856/technom.814.0_251_3

CiNii Research

researchmap

Language：Japanese   Publisher：公益社団法人日本船舶海洋工学会  

CiNii Research

researchmap

Language：Japanese  

CiNii Research

J-GLOBAL

researchmap

Other Link： http://id.ndl.go.jp/bib/3634265

Language：Japanese   Publisher：公益社団法人 日本船舶海洋工学会  

DOI： 10.14856/technom.796.0_755_2

CiNii Research

researchmap

Language：Japanese  

CiNii Research

J-GLOBAL

researchmap

Other Link： http://id.ndl.go.jp/bib/3848893

Language：Japanese  

CiNii Research

J-GLOBAL

researchmap

Other Link： http://id.ndl.go.jp/bib/3848892

Language：Japanese   Publisher：公益社団法人 日本船舶海洋工学会  

DOI： 10.14856/technom.772.0_702_2

CiNii Research

researchmap

Event date： 2023

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2023.html#KawaiK23

Event date： 2023

researchmap

Event date： 2023

researchmap

Event date： 2023

researchmap

Event date： 2022

Language：Japanese  

In this study, we propose a route planning method using floor maps such as facility information maps as prior information. The floor map does not provide geometric information because the details are omitted and in some cases the ratio of height and width is inaccurate. Therefore, we extract the connection relations of the passages from the floor map as a node graph, and calculate the shortest path from the specified current location to the destination based on the node graph.We use LiDAR to detect the direction and branching of the corridor in the real environment, and then follow the calculated path.

researchmap

Event date： 2022

Language：Japanese  

Autonomous mobile robots are required to operate in a wider area of environments. When having a security robot monitor people in the environment, we thought that the robot could be operated more efficiently if it could understand the movement paths of people in the environment. In this paper, we propose a new method for obtaining movement paths information over a wide area. The presence/absence of people in the observable area, the walking speed, and the elapsed time outside the observable area are integrated to estimate the movement paths of people. We can obtain global information of movement paths from local observation information. Through experiments with security robots, we show that it is possible to improve the quality of security by obtaining trends related to human movement in the environment from estimated path information and using these trends in the robot’s patrolling strategy.

researchmap

Event date： 2022

Language：Japanese  

For autonomous mobile robots to act in society, it is important to predict the action of dynamic obstacles. In this paper, we use state representation learning, a form of deep learning, to obtain a predictive model from 2D-LiDAR and the robot’s action that considers the interaction between the robot and dynamic obstacles. This predicts the future for possible actions taken by the robot. The prediction results are compared with GRU, a form of representation learning. We also describe how to use the prediction results for conventional motion planning. As for as we know, there are few cases where prediction results are reflected in motion planning.

researchmap

Event date： 2022

Language：Japanese  

In this paper, we use an RGB-D camera to estimate and track the centroid of a person. In order for a robot to run safely in the crowd, it is important to recognize and track people with high accuracy, and the RGB-D camera is suitable for human tracking because it can not only recognize people with high accuracy based on images but also obtain distance information of the recognized people. The proposed method improves the performance of centroid estimation and tracking by removing noise from the distance information and considering the motion model of the person. The effectiveness of the proposed method is demonstrated by comparing the estimated position and labeling of a person with the true values in an experiment of simultaneous tracking of multiple persons.

researchmap

Event date： 2022

Language：Japanese  

In this paper, we propose a cooperative localization method for multi － robots based on the confidence evaluation of positioning. In conventional methods, robots are often separated into two groups: parent robots and child robots. Since child robots depend on the information of parent robots, the error of parent robots will spread to the whole robot system. Therefore, we propose a method to estimate the position by dynamically switching the reference robots without specifying a specific parent robot. Results show that our method can maintain the accuracy of position estimation for the entire robot system better than when parent robots are fixed.

researchmap

Event date： 2022

Language：Japanese  

With the labor shortage, autonomous mobile robots are required to perform a variety of tasks. When we consider having an autonomous mobile robot perform tasks such as cleaning, maintenance, and inspection, it is assumed that the tasks will change depending on its own position. Therefore, it is difficult to perform the task effectively unless the robot knows its own position and posture with higher accuracy than in the case of simply moving. In this study, we are developing a robot that can move in all directions using a mecanum wheel. And also we are developping an autonomous movement system using a local path planning method that can determine the movement direction and posture independently to make effective use of the omni-directional movement mechanism, and a global path determination method that assumes that a non-expert person can set the path.

researchmap

Event date： 2022

researchmap

Event date： 2022

researchmap

Event date： 2022

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2022.html#KawaiYHSMK22

Event date： 2022

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2022.html#HirakawaK22

Event date： 2021

Language：Japanese  

In this paper, we propose a moving object point cloud detection method using point cloud clustering. Conventionally, point clouds obtained with 3D-LiDAR are placed in discretized spaces called occupancy grids, and the static and dynamic environments are estimated by determining whether points exist in each space over time. However, this method misrecognizes parts of the object as partially static, because parts such as the back legs of a pedestrian tend to remain and continue in the occupied grid. Therefore, we overcome the conventional problem by clustering the point clouds obtained with 3D-LiDAR using Depth Map and making static and dynamic judgments for each cluster based on the estimation results of the occupancy grid.

researchmap

Event date： 2021

Language：Japanese  

Differ from extracting feature description such as SIFT and SURF, semantic segmentation has strong robustness to optical changes such as cross-seasonal and day-night changes. In this paper, we propose advanced visual localization method using semantic segmented images and a mesh map with semantic information made from annotated LiDAR scan data and have solved the problems of previous study. In localization phase, we use traditional Monte-Carlo Localization and calculate likelihood by comparing segmented image from on-board camera and an image of the mesh map landscape as seen from a possible predicted location. This method achieved practical localization accuracy with keeping the benefit of semantic segmentation. A source code used in this experiment is available at following github page. github.com/amslabtech/semantic mesh localization

researchmap

Event date： 2021

Language：Japanese  

In this paper, we propose a new localization method combining visual place recognition with MCL. It is difficult for conventional localization methods to localize in the absence of prior position information. This is especially true in environments with poor geometric features because most of them rely on only geometric information when they localize. Therefore we reflected the results of similarity-based visual place recognition to MCL. MCL is a probabilistic and geometric localization method using LiDAR. An experiment was conducted in a simulation environment to compare MCL using only geometric information and the proposed method using both geometric and visual information. The experimental results show that our method is superior in terms of accuracy and convergence speed.

researchmap

Event date： 2021

Language：Japanese  

In this paper, we propose a jerk-suppressed action planning method for mobility robots or transfer robots using deep reinforcement learning by considering human-to-human and human-to-robot interactions. In a dynamic environment such as a station or an airport, surrounding situations are complex, so there is also a situation where a rule-based planning has difficulties in dealing with. We aim for safe and smooth action planning by using deep reinforcement. In this paper, we propose a jerk-suppressed action planning method for mobility robots or transfer robots using deep reinforcement learning by considering human-to-human and human-to-robot interactions. In a dynamic environment such as a station or an airport, surrounding situations are complex, so there is also a situation where a rule-based planning has difficulties in dealing with. We aim for safe and smooth action planning by using deep reinforcement.

researchmap

Event date： 2021

Language：Japanese  

The purpose of this research is to evaluate the difficulty of an environment that we want to autonomous mobilities to drive only from prior information. We think that to be enable to evaluate the difficulty would contribute to reducing the risk and the number of times autonomous mobilities are required to perform actual driving tests. We propose a method to evaluate the difficulty by using the achievement rate of the mobile task. This rate is calculated by a function with environmental factors as variables that affect autonomous driving. Therefore, we enable to evaluate the difficulty only from prior information. We also analyzed which factors influenced this difficulty by examining the magnitude of each factor’s effect on the estimation of the mobile task achievement rate.

researchmap

Event date： 2021

researchmap

Event date： 2021

researchmap

Event date： 2021

researchmap

Event date： 2021

researchmap

Event date： 2021

researchmap

Event date： 2021

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2021.html#OzakiK21

Event date： 2021

Language：Japanese  

This paper describes the online activities of Tsukuba Challenge 2020. Due to the spread of coronavirus (COVID-19 pandemic), the experiments where all teams gathered in Tsukuba were canceled. As an alternative, we proposed the following activities: navigation experiments at each team’s local site, provide and use of datasets, and use and development of simulators. In addition, to keep the community connections, an orientation meeting, lightning talks, and a symposium were held online.

researchmap

Event date： 2020.1

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2020.html#NagaiKK20

Event date： 2020

Language：Japanese  

This paper presents a survey of the Tsukuba Challenge 2019. The Tsukuba Challenge is an experiment for autonomous navigation of mobile robots on public walkways. Pedestrian paths through the city are given as the navigation tasks. Participating teams develop their own robot hardware and software. We investigated the results of the real world experiment for all teams.

researchmap

Event date： 2020

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2020.html#OzakiK20

Event date： 2020

researchmap

Event date： 2020

researchmap

Event date： 2020

researchmap

Event date： 2020

Language：Japanese  

In this paper, we propose a trajectory prediction method in an environment where pedestrians are surrounded by obstacles. When an autonomous mobile robot runs around a pedestrian, it is necessary to avoid the pedestrian safely. For that purpose, it is important to predict the trajectory of the pedestrian and take the avoidance action an earlier stage. In this method, the LSTM encoder-decoder model is used to predict the complex behavior of pedestrians. The input of the model is the obstacle information around pedestrian obtained by range-finder sensor and the past position of the pedestrian. Through experiments in a simulation environment, we show that the proposed model can predict the behavior of pedestrians affected by surrounding obstacles.

researchmap

Event date： 2020

researchmap

Event date： 2020

Language：Japanese  

In this paper, we proposed learning-based motion planner for autonomous mobile robot which outputs discrete motion commands by referring from the onboard RGB image and target position without using map. It is difficult to get policy for navigation of mobile robots using only visual sensor. In recent years, research on visual navigation using deep neural networks has been conducted. However, these methods do not have a structure to store time-series information in a network. Therefore, this paper presents the DNN motion planner with LSTM trained by using expert trajectories. As a result, we show that our method can navigate agent in simulation environment without map.

researchmap

Event date： 2020

researchmap

Event date： 2020

researchmap

Event date： 2020

researchmap

Event date： 2019.4

Language：English  

researchmap

Other Link： https://dblp.org/db/conf/sii/sii2019.html#SadakuniKOK19

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

Language：Japanese  

In this paper, we propose a learning-based motion planner for autonomous mobile robot which output continuous motion commands by referring from the scan data of 2D-LiDAR and the target point without using environmental map. We aim to use this method for a safe and efficient navigation system of an autonomous mobile robot with high generalization performance in urban environment. We show that a mapless motion planner can be trained through a deep reinforcement learning method in only simulator. In order to verify the effectiveness of the proposed method, collision avoidance and navigation performance are evaluated by directly applying the learned planner to the actual machine in the real world. As a result, we showed that we can obtain navigation performance equivalent to conventional method without using pre-environment map.

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

Language：Japanese  

This paper presents a pose estimation method for mobile robots with vertical walls in urban areas. This proposed method exploits the fact that most of all artificial walls are built vertically. It corrects accumulative error of estimated poses by estimating the gravity vector with the fact without any 3d maps or 3d models. The poses which are estimated with vertical walls are integrated with integration of angle velocity from inertial sensors thorough extended Kalman filter. To evaluate the proposed method, outdoor experiments with an actual robot are performed. It shows the method keeps correcting accumulative error while the robot moves.

researchmap

Event date： 2019

Language：Japanese  

When a mobile robot moves in dynamic environment, estimating collision with dynamic obstacles and avoiding collision in limitation of mobility of the robot are required. In this paper, we address differential drive robot which is most popular driving mechanism for mobile robots, and four wheel drive-steering robot which is popular as a driving system for mobile manipulators but not popular as a driving system for mobile robots. We think that since four wheel drive-steering robots can change their direction by steering, they are able to avoid obstacles more agilely than differential drive robots. We evaluate the two kinds of driving mechanism in their mobility.

researchmap

Event date： 2019

researchmap

Event date： 2019

Language：Japanese  

It is very important for the autonomous mobile robot to recognize the environment. We have developed Wide FOV 3D LiDAR aiming for better environmental recognition than general 3D LiDAR. Wide FOV 3D LiDAR is far superior to the vertical viewing angle of the conventional one. Furthermore, its rotational scanning speed is controllable. Therefore, if the rotational scanning of Wide FOV 3D LiDAR is slowed down, its point cloud becomes high density. However, the slow rotational scanning will lower real-time performance of 3-dimensional scan. This paper presents the density control method of the point cloud with Wide FOV 3D LiDAR for compatibility between generation of high density point cloud and maintaining of real time performance of 3-dimensional scan. Our proposed method extended object detectable distance without reducing real-time performance.

researchmap

Event date： 2019

Language：Japanese  

In this paper, we propose a human detection method triggered by a visual sensor. In order for moblile robots to work in urban areas, it is important to be thoughtful for people. And they need to recognize people to a certain distance. 3D LiDAR Now widely used has a narrow viewing angle and cannot recognizing nearby obstacles such as curbs. With the wide FoV 3D LiDAR we use, nearby object can be detected, but instead it is not good at recognize distant objects. In this method, we detect a person first by vision sensor, and the position of the person is detected by scanning with 3D LiDAR. The proposed method is shown to enable detecting more 2m distant person than traditional method.

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2019

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

Language：Japanese  

The localization for automobile robot is one of the essential elements. The method for it uses registration using the surrounding geometric shapes by getting the point cloud from LIDAR. So, the estimation of the exact position using traditional methods is difficult in flat areas. This paper presents the method for robot localization using intensity of the ground. By getting features from differences in colors of road marker and materials such as grass and asphalt, our proposed method realizes exact registrations at environments with poor geometric features. The values of the intensity in the same place changes depending on the time and weather. The problem of changing values is solved by applying SLAM algorithm. SLAM algorithm locates itself by getting the point cloud at same conditions. Our proposed system estimates the exact position using the proposed method at environments with poor geometric features and use the traditional method at other places.

researchmap

Event date： 2018

Language：Japanese  

In recent years, the development of robots for substituting and supporting human work in an environment, where we live is actively performed. These robots need to behave with high versatility. So, a method called machine learning is effective to experience these behaviors. For robots to learn flexible motions like human, a method called deep reinforcement learning is generally used. But learning complex motions requires a lot of time, and improvement of learning efficiency is necessary. Therefore, we propose a method which combines deep reinforcement learning and imitation learning that imitates the behavior of the experts. Manipulators acquire a certain level of motion by imitation learning, and then refine its motion by deep reinforcement learning. The process makes learning complex motions efficiently. Experiments of opening a door with a manipulator shows the usefulness of this method.

researchmap

Event date： 2018

Language：Japanese  

Traveling in an urban environment, the autonomous mobile robot needs to recognize the surrounding environment in detail. As a conventional method, an obstacle detection method using the 3D point cloud given by LiDAR can be cited. The 3D point cloud given by LiDAR can acquire detailed position information, but it si difficult to recognize what the detected obstacle is. In recent years, object detection using deep learning has attracted attention, leaving detection accuracy exceeding the model base method. However, depth information using a stereo camera is unstable depending on the environment. In this research, we propose a real-time three-dimensional object position detection method by integrating object detection information obtained from a camera and three-dimensional information acquired from 3D-LiDAR.

researchmap

Event date： 2018

Language：Japanese  

In this paper, we propose a localization system using image comparison. In the conventional localization method used with a 3D-map, it takes a lot of cost to collect prior information. However, by only using the image and the coordinates of the taken location are used as the prior information in our method. Therefore, not only can you collect prior information by running the robot, but you can also use the existing large-scale data. This makes it possible to collect prior information more efficiently. In this system, real-time images are reconstructed as a threedimensional point cloud. The robot will estimate the relative relationship between the prior images, real-time images, and the point cloud, by matching these images and realize its location.

researchmap

Event date： 2018

Language：Japanese  

Conventionally, the destination of an autonomous mobile robot using the map of the running environment has been given using the coordinates in the map coordinate system. In the future, in order to be active in the human society, it is desirable to be able to specify the destination with a specific name as how human will do. As a previous study, a method of giving information to the created geometric map has been proposed, but there are problems in terms of efficiency such as necessity of running in the same environment multiple times. Therefore, in this research, we propose a method to automatically create a map with information related to the name of the environment by integrating information obtained from the camera and LIDAR mounted on the robot.

researchmap

Event date： 2018

Language：Japanese  

In recent years, as the place of active autonomous mobile robot moves to human society, it is necessary to operate while anticipating the movement of a person in a situation such as a crowd. Conventionally, this prediction is often performed under the assumption that the speed of the target object is calculated, and the speed is maintained and acted. Therefore, it is impossible to respond to actions with changing speed from the relationship with the surroundings such as a person passing by. Therefore, in this study, motion prediction of dynamic obstacles is performed using 3D-CNN. 3D-CNN is used for time series 3D point cloud to extract shape information and time change. Based on the relationship with the surroundings of each extracted information, we estimate the future position of each point cloud. By doing this, it is possible to predict motion of the object in a situation difficult with the conventional method.

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Event date： 2018

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Event date： 2018

researchmap

Other Link： https://dblp.org/rec/conf/syroco/2018

Event date： 2018

researchmap

Event date： 2017

researchmap

Event date： 2017

Language：Japanese  

This paper discusses transformation control of an autonomous mobile robot with multiple running modes. For an autonomous mobile robot to move in a human living environment, especially a place like a crowd, smooth action like a two-wheel mobile robot is required. However, with a two-wheel mobile robot, attitude control is required even at rest, and there is a problem that stability is lacking. We propose a method that enables transformation from four wheels to two wheels to obtain stability during rest in addition to smooth running in two wheels mode. Calculate the target attitude angle from the equilibrium of the moments, and transform it by controlling the attitude of the robot upper body and the wheel. In order to stabilize the transformation, feedback control is performed from the difference between the attitude angle obtained from the IMU and the target attitude angle.

researchmap

Event date： 2017

Language：Japanese  

Machine learning attracts attention in environment recognitions. Annotations of an enormous dataset are mainly performed with many time and effort. Therefore, the system of an automatic annotation of the dataset is required. We propose the dataset construction system by the machine learning with a cloud server. In our approach, image collection is performed by multiple robots with a cloud server, and automatic annotation is done by machine learning. In addition, misrecognitions of the machine learning are corrected by hand, and classification re-learns the corrected results. Finally, we compared the required time, that dataset construction by our approach took, with that by previous methods.

researchmap

Event date： 2017

Language：Japanese  

A map integrating approach is proposed in this paper for solving the problem of labor saving. When constructing a large-scale map using multiple robots, the coordinate system of each maps created, must be unified into one coordinate system. Since the area where the robot is active has a large scale, both outdoors and indoors, a map building method which can correspond to any environment is necessary. We propose a method to integrate plural maps, each having different coordinate systems by using informations such as combining landmarks and road shapes. An optimization process is performed on a map created by multiple robots, thereby realizing constructions of a map with a higher accuracy. In order to show the usefulness of this method, we will conduct a map building experiment in the area, which includes indoors.

researchmap

Event date： 2017

Language：Japanese  

Along with the aging society, electric wheelchair have been popular. Accident caused by carelessness or misoperation have increased as electric wheelchair became more popular. To prevent accidents, drive assistance systems are being developed. These are two problems:falling accidents due to undetected steps and path planning not considering the driver's intention. We developed a driving assist system that consists of step doundary detection using 3D-LiDAR and path plans considering the driver's intention. In order to detect a travelable area, our system is the integration of evaluation suitable for each factors that make the traveling hard. The evaluations are composed of the height information and the curvature estimation. We generate trajectory candidates that can safely travel from detected results. Afterwards, the electric wheelchair selects and follows the trajectory considering the driver's intention.

researchmap

Event date： 2017

Language：Japanese  

Many 3D-LiDAR‘s that are currently being made has been developed for autonomous cars. When an autonomous mobile robot moving in urban environments uses the 3D-LiDAR, two problems occur. One is that it is difficult to recognize near objects. The other is high power consumption. To solve these problems, it is necessary to widen the viewing angle in order to recognize the object at a short distance, reduce the distance measuring point in order to lower the power consumption, and control the point density for any area. In this paper, we will develop a full field-of-view 3D-LiDAR with wide view angles of omnidirectional and vertical 97.8 degrees and controllable point-density in any area from near to far.

researchmap

Event date： 2017

researchmap

Event date： 2017

researchmap

Event date： 2017

researchmap

Event date： 2017

researchmap

Event date： 2017

researchmap

Event date： 2017

researchmap

Event date： 2016

Language：Japanese  

Reinforcement Learning can allow robots to learn a policy, which is a rule of choosing optimal behaviors in unknown environments. While it is effective for simple tasks with small action-state space, it cannot deal with mobile manipulators due to large action-state space. In this paper, we propose a method that is able to learn a policy of a mobile manipulator using Deep Reinforcement Learning. This method consists of Q-Learning, which is representative algorithm of Reinforcement Learning, and Deep Learning. To show the effectiveness of our method, we first train the system with a policy by simulation. After the training, the system achieves to press the elevator buttons in real environments.

researchmap

Event date： 2016

Language：Japanese  

This paper describes the design, system of the controllable point density and evaluation of a 3D-LIDAR. 3D-LIDAR became popular to autonomous robot because of its accuracy of recognition and the ability to obtain 3D-points of surrounding environment. However, previous 3D-LIDAR has narrow vertical view angle that is difficult to detect close 3D-object such as human. In addition, the more distance object is, the more sparse the points will be, so it will be difficult to detect distant object accurately. Our 3D-LIDAR consists of 3 titled 2D-LIDAR on the table. By spinning this device, we could manage to scan data from 98 degree vertical view angle. Furthermore, our 3D-LIDAR could heighten point density in any area to change rotation speed that is able to detect more distance object.

researchmap

Event date： 2016

researchmap

Event date： 2016

Language：Japanese  

Scan matching is a general method which estimates accurate pose of robots by comparing sensor data with prior information. However, this method is not enough to cope with the environments which consists of regular patterns (e.g. symmetry, point symmetry). Furthermore, it takes a long time to estimate pose in huge area where robots move. To solve these problems, we combine the scan matching and landmark recognition. In our approach, the pose of a robot is represented as probabilistic distribution, and the probability around the landmark is increased by the recognition. This processing better estimation in the regular-patterned environments. In addition, the process reduces the time required for position estimating.

researchmap

Event date： 2016

researchmap

Event date： 2015

Language：Japanese  

In this paper, we propose a driving assist system for electric wheelchair. In several countries as the rate of senior citizens increases, the number of electric wheelchair users and the accidents by the senior becomes larger. To decrease these numbers, the wheelchair can avoid or stop before the confliction with the obstacles, is needed. Using the pointcloud from tilted 2D LiDAR, we construct the environment map around own position and generate the practicable paths. Considering the driver's command, safety, and smoothness, the path is selected and followed finally. To prove the validity of our method, we ride on the wheelchair and drive the several environment in our campus.

researchmap

Event date： 2015

researchmap

Event date： 2015

Language：Japanese  

In this paper, we propose a elevator detection independent from map data. The elevator generally have a lot of designs such as the simple in the station and the decorated in the department store. It seems that the elevator has few common features, but door and button are equipped with the all ones. The former has a certain width and depth between the wall, the latter is the distinctive shape (the plate and the button). Thus, we recognize these two features using by the LIDAR data and the camera image. The algorithm achieves 90% detection without the map, which demonstrates that our method is generic and robust with variations of color, texture, lighting condition.

researchmap

Event date： 2015

Language：Japanese  

This paper describes autonomous mobile robots navigation system and automatic large-scale 3D-map construction method. In previous works, most of them are using detailed map, which is created by operator in advance. Prior detailed map is a very effective tool to improve working efficiency, stability of the robot, but increasing the workload of operator is undesirable to generalize the robot in the real world. Furthermore, autonomous mobile robots have to function appropriately even in areas where they have never been before. Thus pre-driving details are not needed in our system. LIDAR is mainly used in our system, in creating local map, avoiding the moving object (Motion Planning), and recognizing own position (Localization) and environment around the robot (Environment Recognition). Moreover, we construct the detailed map automatically by using the autonomous robot driving data to improve accuracy of the navigation system without increasing the workload of operator. The availability of proposed method is confirmed by the experiments in Meiji University.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose a teleoperation system supported by an autonomous mobile technique. The traveling direction of the robot is roughly determined by the operator, and that motion planning is executed to achieve a safe driving. In the motion planning phase, multiple trajectories that take into account the mobility of the robot are generated, and the most optimal trajectory is singled out through several evaluations. These evaluations consider a risk of collision, a smoothness of the trajectory, and a difference between traveling directions and terminal states of trajectories. The risk of collision is checked by using 2D navigation map obtained from the laser scanned data. By the above-mentioned technique, we construct a teleoperation system which attains safe driving and desired motion of the robot. The effectiveness of the proposed method is proved through the experiment in complex environments.

researchmap

Event date： 2014

Language：Japanese  

JAXA is planning to launch Hayabusa 2 in December 2014. Hayabusa 2 will be equipped with an asteroid exploration rover which is developed by the consortium of universities. A goal of the rover is establishment of the technology for locomotion on asteroid surface. In our previous work, we proposed a new moving mechanism called "Environment-Driven Torquer, EDT." The EDT is driven by environment temperature with bimetal. In other words, the EDT can move the rover without both battery and CPU. We have developed an engineering model with magnet latch. In this paper, performance evaluation of the EDT is executed by free-fall experiments in ZARM.

researchmap

Event date： 2014

Language：Japanese  

When we use odometry for localization of the robot, accumulated error which caused slip of tire become the problem. Currently, by the spread of the internet, map information is easily obtained. Therefor we can obtained priori information such as building and roads without go directly to the place. In this paper we propose the method which reduce the accumulated error by matching outline information of the building that was extracted from digital map and wall surface information of the building that is obtained from LIDAR. Our experiment was performed over 1000 m trajectory approximately in Meiji University to verify its validity.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose a human recognition for the autonomous mobile robot using LIDAR. For traveling the autonomous mobile robot, need to recognize of the surrounding environment. In particular, grasp to the humans is important in a railway station or the like. Generally, the camera image is used for the recognition of the human, but it is susceptible to the influence of a change in brightness. In our method, performing recognition using a three dimensional LIDAR, and to recognize the human in a wider range. Furthermore, to perform a stable and high accuracy recognition by learning shape features using SVM. Our experiments are performed indoor and outdoor environments that contain a Tsukuba city, Meiji University to verify its validity.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose a method for travelable area detection using a 3DLIDAR. In order for the autonomous mobile robot to travel safely, travelable area detection is necessary in such areas where traveling becomes hard by differences in the height and material changes. Our system is integration of evaluations suitable for each such factors that make traveling hard. The evaluations are composed of the height information, the curvature estimation, the classification of the reflection intensity. Afterwards, we construct a local environment map of the robot from detected results. The effectiveness of the proposed method is proved through the experiments results in outdoor environments.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose an azimuth estimation method for mobile robots. In our method, artificial walls within the detection range of the LIDAR mounted on the robot are used for estimation of the azimuth angle. Each wall is approximated as a plane, and the normal vector of this is used for some processes. By means of this reasonable approximation, the surrounding environment is described with extremely little amount of data. This leads the real-time property of the method. Additionally, taking account of the kinematic performance of the robot, robustness and high-responsiveness is realized. At the end of this paper, the utility of our method is shown through two experiments.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose a method of judging status of light without previous light position data. Our system makes list of light bulbs registered position and status of their while running with detecting light bulbs, estimating position of their composed with 3D-model of surrounding environment. When robot patrol and check light bulbs, comparing camera image and list's data and renewing light bulb list with estimating process. Experiments performed accuracy and effectiveness of the proposed method in indoor environment.

researchmap

Event date： 2014

Language：Japanese  

Asteroid probe Hayabusa 2 is developed by JAXA. Asteroid exploration rovers will be carried by Hayabusa 2. In our research project, we have developed an environmental-temperature-driven buckling actuator with a bimetal thin plate for a hopping mechanism of the rover. In this study, to characterize the hopping capability of the actuator, we conducted microgravity examination in ZARM drop tower. We fabricate a test rover for the microgravity examination with a rapid heating system for actuation in short free-fall time. Although reaction force from baseplate was incomplete in take-off, the rover hopped with a vertical velocity of 3.12 mm/s with 0.035 rad/s rotating rate.

researchmap

Event date： 2014

Language：Japanese  

In this paper we focus on the problem of Visual Odometry. In recent years, several VO algorithms have been SAD or ZNCC in stereo matching and feature points tracking. These methods are template matching with luminosity distribution and have properties of accuracy of pixel and poor change of lightness of the image. These properties cause low accuracy of correspondence of feature points. In this work we employ Phase-Only Correlation in stereo matching and feature points tracking. POC only use phase of image and have the properties of high accuracy of subpixel and strong in change of lightness. We experiment with image data set of rough terrain and experiment result show that this method provides the high accuracy of VO.

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose communication and observation system based on web for remote control robot. For observation by remote control robot, need a system to perform stable communication. Using the communication by radio techniques, there is maybe no way to control the robot from distant place such as several hundreds of kilometer away or several thousands of kilometers away. On the other hand, using web techniques, it also have the problem, it is difficult to communicate data in same time between robots and operator. Data are often broken because of failure of synchronization. In our method, we construct the communication system based on web, and use the new technique of web-communication called websocket. Websocket improve the communication speed and prevent data from being broken. The effectiveness of the proposed method is proved through experiments of observation by robot controlled from distant place.

researchmap

Event date： 2014

researchmap

Event date： 2014

Language：Japanese  

In this paper, we propose a indoor environment mapping method for large space with multi-robot. We build a large space map with map fusion by the maps that multi-robot built. Our system is composed of 2DLaserSLAM, the mutual recognition with the camera and map fusion with ICP. When moving robots recognized it mutually, we integrate maps formed from each other's 2DLaserSLAM using ICP. The effectiveness of the proposed method is proved through experiments in indoor environments.

researchmap

Event date： 2013

Language：English  

researchmap

Event date： 2013

Language：Japanese  

In this paper, we propose an autonomous navigation system that do not need detailed preinformation. To achieve robust traversability, our system adopts following technologies. The system involves a localization with GPS and dead reckoning. We integrate it by using Divided Difference Filter (DDF). Additionally our system also estimates traversability of road regions using a laser range finder (LRF). Our local map is made in consideration of pedestrians by a moving obstacles detection. The effectiveness of the proposed system is proved through some experiments.

researchmap

Event date： 2013

Language：Japanese  

Volcano observation rover needs to be able to explore and observation at the place which man cannot come into for a long period of time. Therefore, the Volcano observation rover must be carried out in the independent energy supply. Study of Volcano observation rover, observation of the volcano have been made in the world. However, there is no robot has been active robot is a long period of time to replenish energy at independence. In this research, We propose a power system that can be long-term observation. In order to supply electric source and wind generation by itself, the rover carried the solar panel. The power system is equipped with four modes. By selectively using these modes of power availability, allowing the longterm observation. It can carry out experiment for three consecutive days. In the developed system, perform the experiment.

researchmap

Event date： 2013

Language：Japanese  

Mobile robots have to recognize own current position. In general, Global Positioning System(GPS) have been used to know its position, but many kinds of surroundings usually give rise to significant positioning error. Therefore we have been proposed place recognitions method using images instead of GPS. However feature-based image retrieval system could not recognize where feature could not be extracted. In this research, we propose a place recognition method using Bag of Regions. Images, which are taken from reference course, are divided into segmentations by Simple Linear Iterative Clustering (SLIC). These segmented regions are described by intrinsic parameters such as color space histogram and represented as "regions" by Bag of Regions. Regions make possible image matching even in an environment where it is difficult to take interest points such as Harris corner, Fast Hessian and so on. To calculate matching probability between current obtained image and reference images, we employ Fast Appearance Based MAP(FAB-MAP) scheme. Image matching can be more robust to disturbances by taking into account the probability of co-occurrence of those words. To verify the validity of proposed method, our experiment is conducted course include indoor and outdoor environments.

researchmap

Event date： 2013

Language：Japanese  

In this paper, we propose a method for localization accuracy improvement by reflecting evaluation of reliability of information from GPS in outdoor environment. Currently,GPS is widely used to obtain the positioning information in urban environment. However,if there are obstacles such as trees and tall buildings around the robot, signal lost and multipath is occur. Therefore, all the obtained position information is not necessarily correct. Consequently, processing each threshold with respect to the Satellites number ,positioning techniques and Carrier Noise Ratio. Additionally, to evaluate the likelihood of GPS observations by caluculating the normalized innovation square by using the motion model,we will get only the more reliable GPS observations. We show robust localization even if large error occurs in GPS observation by reflecting result of the evaluation to the localization

researchmap

Event date： 2013

Language：Japanese  

In this paper, we propose a method for travelable area detection in an urban environments using a LRF. Non-travel in the urban environment can be divided into roughly, change in the height and different materials. Our method using a laser range finder to get the reflection intensity and height information of the obstacles. The system recognizes using appropriate methods to observations obtained. Non-travel detection is edge points detection by the second order derivative using height information, and the classification by the machine learning using the reflection intensity. Afterwards, we construct a local environment map of the robot from detected results. The effectiveness of the proposed method is proved through the experiments results in outdoor environments.

researchmap

Event date： 2013

researchmap

Event date： 2013

Language：Japanese  

If you want to explore the unknown environment using an autonomous mobile robot, it is necessary to estimate the position and orientation of the robot relative to the environment to be searched. We also require a high-resolution three-dimensional map of the environment exploration. When we use the 2DLaserSLAM that depends on odometry, it must be considered that exploration environment is assumed to be two-dimensional plane. But such an environment is limited. Because it requires the odometry, an expensive gyro sensor, an acceleration sensor or the like is required in the mobile does not have wheels. In this study, we propose a 3DLaserSLAM using 3DLRF only as an external sensor. We show that without choosing the mobile environment, exploration, the road surface by the proposed method, the construction of the three-dimensional map high-definition can be run on a single sensor, and pose estimation self-position in 3D space.

researchmap

Event date： 2013

Language：Japanese  

When signs of volcanic activity are shown, it becomes difficult to conduct scientific investigation near the volcano for a long term. We developed the rover named "Micro6 Volcano" that can explorer active volcanoes without external battery for a long term, and has autonomous mobile capacity. Our long term investigation which took approximately two weeks was conducted at the middle of Mt.Mihara in Izu Oshima under the assumption that the eruption crisis of the volcanoes looms. During this term, Micro6 Volcano observed meteorological information by remote operation. It has communication device using mobile phone line to send images and observational data such as locational data and so on. We investigated by controlling the rover remotely with consideration for these data at Meiji Univ. in Ikuta, Kawasaki where away from Mt.Mihara about 100 km. This investigation succeeded in moving observation over hundreds of meters and operation through some nights with saving power mode.

researchmap

Event date： 2013

researchmap

Event date： 2013

Language：Japanese  

In this paper, we propose a motion planning method for reaching goal-point with target heading attained. Our system is composed of global planning and trajectory generation. The generated trajectory is considered the robot motion model. Candidate trajectory is generated plural,and the optimal trajectory is singled out through a three-stage evaluation. The effectiveness of the proposed method is proved through experiments in outdoor environments.

researchmap

Event date： 2012

researchmap

Event date： 2012

researchmap

Event date： 2012

Language：English  

researchmap

Event date： 2012

researchmap

Event date： 2012

researchmap

Event date： 2012

researchmap

Event date： 2012

researchmap

Event date： 2012

researchmap

Event date： 2012

Language：Japanese  

In this paper, we propose a method for detecting and tracking moving obstacles in an outdoor environment. Our system clusters ranges observed by a laser range-finder. Moving obstacles are not estimated from clusters directly, we assume clusters as candidates of moving obstacles. The system recognizes moving obstacles in the candidates to track them with Extended Kalman Filter. The effectiveness of the proposed method is proved through experiments in outdoor environments.

researchmap

Event date： 2012

Language：Japanese  

In this paper, we propose a mobile robot autonomous navigation system using Theta^*. Our system uses traversability map constructions and Theta^*, a variant of A^*. In order to analyze traversability, a plane is estimated from point cloud which comes to be sparse in proportion to distance from stereo camera. Traversability is analyzed from height, slope and roughness of the plane. Path planning is performed using Theta^* on the grid map was obtained. Grid map are often used to represent terrain. However, paths formed by grid edges can be sub-optimal and unrealistic looking, since the possible headings are artificially constrained. Theta^* propagates information along grid edges without constraining the paths to grid edges. Thus, Theta^* finds shorter and more realistic looking paths. The effectiveness of the proposed method is proved through experiments in rough terrain.

researchmap

Event date： 2012

Language：Japanese  

This paper has proposed the method of merging the map that multi mobile robots built, respectively. Until now, in order to merge maps, map data needed to overlap at least 80%. Then, accurate map merging is enabled by presuming the data with which it is not overlapped from map data, and extracting from merging process. Thereby, even 45% of duplication enabled map merging, and realized time shortening larger than the conventional technique.

researchmap

Event date： 2012

Language：Japanese  

In this research, we propose a mobile robot localization based on GPS error evaluation in urban environments. Localization in urban environment has still a challenging problems such as multipath and signal lost. Our proposed method uses a localization with GPS which is evaluated the reliability of GPS observations, dead reckoning and integration method using Divided Difference Filter (DDF). Proposed evaluation algorithm uses satellites elevation and azimuth that are leveraged in global positioning system in order to calculate covariance matrix. Especially, our evaluation focus on calculated real-time deviation from covariance matrix. The covariance matrix is associated with DDF. Hence, the proposed system makes mobile robot localization less likely to be affected by accuracy of GPS observations. The effectiveness of the proposed method is proved through experiments in urban environments.

researchmap

Event date： 2012

Language：Japanese  

In this paper, we propose a mobile robot localization system using visual place recognition. Our system uses visual place recognition and dead reckoning using wheel odometry. We use the theory of the observation likelihood estimation introduced from Fast Appearance Based Mapping (FAB-MAP) for visual place recognition. Images collected by a robot produce Bag-of-Words representation and Chow Liu trees. This representation is based on quantized SURF detectors/descriptors. A certain place can be recognized when the collected image matches a reference image coming from the same place in reference scene database. In order to solve problems of appearance based place recognition, we use sensor fusion using motion model. In addition, our system use road pattern matching between robot view and reference scene from visual place recognition. This enables more accurate localization than previous approach. Experimental results are presented in an outdoor urban environments. The proposed localization system can achieve the average robot's position error within 2m without GPS in an outdoor environment.

researchmap

Event date： 2012

Language：Japanese  

In this paper, we propose an autonomous navigation system. To examine the traversability robustly, the system adopts the following technologies. The system involves a localization with GPS and dead reckoning, and integration method using Divided Difference Filter (DDF). The system also estimates traversability of road regions using a laser range finder (LRF). Lastly, the system detects moving obstacles. The effectiveness of the proposed system is proved through some experiments.

researchmap

Event date： 2012

Language：Japanese  

The authors are involved in the development of a tiny rover system onboard "Hayabusa-2" spacecraft which is scheduled to launch in 2014 for the Near Earth asteroid "1999JU3." The development of Hayabusa-2 started very urgently after its predecessor spacecraft "Hayabusa" made a self-sacrificing reentry into the Earth in 2010. The authors have a heritage how to develop a surface exploration rover through the experience on "MINERVA" rover in Hayabusa mission. But the rover system for Hayabusa-2 is a brand-new one since it is designed to fit the new target asteroid, of which surface environment and conditions are different from those of the previous one. The mobile system of the primary rover has the same strategy on its hopping movement as MINERVA, but is slightly different on the configuration of the actuators. This paper describes the mobile system of the primary rover onboard Hayabusa-2.

researchmap

Event date： 2012

Language：Japanese  

Volcano exploration rover needs to be able to explore at the place which man cannot come into for a long period of time. For that purpose, various technical subjects, such as electric power management, communication, and waterproofing, must be solved. In this research, Volcano exploration rover Micro6 was developed. It can carry out operation in a long period and remoteness. The drive mechanism which enables a stable driving of volcanic irregular place was designed. In order to supply electric source by itself, the rover carried the solar panel. In the developed system, perform the experiment.

researchmap

Event date： 2011

Language：Japanese  

This paper describes a forward velocity estimation method using visual information. Forward velocity estimated from ground surface images is not affected by wheel slip. Using the estimated velocity, estimation error of localization would be decreased. In order to estimate velocity accurately and stably, many feature points are extracted and tracked with ground surface images. Shifting of feature points in an image is transformed into movement distance of body by plane projection for calculating the forward velocity. Experimental results in a rough terrain environment like other planet's surface are presented for verifying our system.

researchmap

Event date： 2011

Language：Japanese  

This paper describes intersection recognition in urban environments to acquire appropriate motion decision for autonomous navigation. Intersection shape which includes width, curvature, position and heading of a road, is given by road geometry scanned from onboard laser sensor. It is proposed that an intersection existence can be classified from constructed map. The proposed method is implemented and the effectiveness of the intersection recognition is verified in outdoor environment.

researchmap

Event date： 2011

Language：Japanese  

researchmap

Event date： 2011

Language：Japanese  

In this paper, we propose a learning method to solve the parameter determination problem of Divided Difference Filter (DDF) for accurate localization. DDF can achieve comparatively accurate localization than other Kalman Filter algorithms in poor GPS area. However, parameter determining process of DDF requires significant time and engineering cost. Furthermore, it is difficult to obtain optimal parameters for accurate localization by hand-tuning. DDF has three parameters which should be determined: covariance matrices of input and measurement noise and a Hyper-parameter. Our technique uses a discriminative learning method to determine these parameters. The proposal method absolves developers from the cumbersome process of parameter setting. This paper describes the efficiency of our technique through simulations and an experiment.

researchmap

Event date： 2011

Language：Japanese  

This paper describes a traversability map construction for autonomous mobile robots. To represent a terrain, a plane is estimated in each grid from point cloud data using RANSAC algorithm. Traversability is analyzed from height, slope and roughness of the plane. Point cloud data becomes sparse in proportion to distance from stereo camera. Thus, point cloud sampling areas are expanded using stereo vision error model. Finally, we have run our rover testbed "Micro6 " in Ura-Sabaku desert of Izu-Oshima island, and experimental results were shown that a proposed method validities.

researchmap

Event date： 2011

Language：Japanese  

In this paper, we propose a speed control system for autonomous mobile robots using long-range road estimation. Our speed control determines velocity from the width and distance of the road surface obtained by the projective transformation of the result of road estimation. Our proposed method adapt to various environments. Moreover, it is effective to moving objects. Experimental results show that efficient movement can be performed by proposed method.

researchmap

Event date： 2011

Language：Japanese  

In this paper, we propose a mobile robot localization system using appearance based place recognition. Our system uses appearance based place recognition and dead reckoning with Wheel Odometry. We use the theory of the observation likelihood estimation introduced from Fast Appearance Based Mapping (FAB-MAP) for appearance based place recognition. Images collected by a robot produce Bag-of-Words representation and Chow Liu trees. This representation is based on quantized SURF detectors/descriptors. A certain place can be recognized when the collected image matches a reference image coming from the same place in scene database. In addition, problems of appearance based place recognition techniques are solved by using sensor fusion. Experimental results are presented in an outdoor environment. The proposed localization system can achieve the average position error within 25m without GPS in an outdoor environment.

researchmap

Event date： 2011

Language：Japanese  

This paper describes a moving object detection method in outdoor environments. Probabilistic grid map is introduced to mitigate the impact of occulusion. Our method applies x-means to cluster discrete data and track to several objects correctly. In order to track objects robustly and estimate state of objects, particle filter is introduced. The approach is tested in an outdoor environment using 2D laser-range-finder. The experimental result indicates that our system is able to track moving objects robustly.

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：English  

researchmap

Event date： 2011

Language：Japanese  

This paper describes design and development of planetary exploration rover ground testbed Micro6. Micro6 haves high precision sensors and general purpose computers so that not only the driving test of an irregular ground but also research and development of software can be done. We select the apparatus carried in Micro6 and design the drive mechanism it can fully drive. A fundamental driving test is done on the volcano surface at Mt. Mihara in Izu-Oshima.

researchmap

Event date： 2011

Language：Japanese  

This paper describes a traversability map construction for autonomous mobile robots. To represent a terrain, a plane is estimated in each grid from point cloud data using RANSAC algorithm. Traversability is analyzed from height, slope and roughness of the plane. Point cloud data becomes sparse in proportion to distance from stereo camera. Thus, point cloud sampling areas are expanded using stereo vision error model. Finally, we have run our rover testbed "Micro6" in Ura-Sabaku desert of Izu-Oshima island, and experimental results were shown that a proposed method validities.

researchmap

Event date： 2011

researchmap

Event date： 2011

researchmap

Event date： 2010.6

Language：English  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Event date： 2010.6

Language：English  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Event date： 2010.6

Language：English  

researchmap

Other Link： https://dblp.org/rec/conf/isr/2010

Event date： 2010.2

Language：English  

researchmap

Event date： 2010.2

Language：English  

researchmap

Event date： 2010

researchmap

Event date： 2010

Language：Japanese  

This paper describes design and development of Micro6 of drive mechanism. Micro6 is a ground examination special-purpose machine of planetary exploration rover. In order to enable Micro6 to a ground run, calculating the torque of a drive part and the optimal design are carried out. The developed examination machine gets over a level difference, a running test of the irregular ground and a turning test. And the validity of this mechanism is shown.

researchmap

Event date： 2010

Language：Japanese  

This paper proposes a platform for autonomous mobile robot which both beginners and researcher can use. We describe to improving development efficiency by offering the environment equipped with the function required for an autonomous mobile robot. And we offer the platform which can be used to share results and replicate each other's experiments. Our platform can solves the complication of a software system, the increase in a function demanded and a burden at the time of learning a robot, and the existing of various standard and specifications. It is shown that this platform is actually used and a complicated autonomous move plan can be performed.

researchmap

Event date： 2010

Language：Japanese  

In this research, optimal path generation which robots can trace easily and safely is proposed. GPS is generally used for path generation in outdoor environments. However GPS has some troubles. It is shown that proposed method can automatically generate consistent path to follow easyly even if a robot is in poor GPS area. The path is optimized by focusing on locally consistency of odometry trajectory and little accumulation error of GPS. In order to optimize the trajectory, evaluation function of Graph SLAM alogorithm is introduced. An experimental result shows effectiveness of our approach.

researchmap

Event date： 2010

Language：Japanese  

In this paper, we propose an automatically learning technique to solve tweaking parameter problems of Unscented Kalman Filter (UKF) for accurate localization. The parameters consist of three kinds of parameters: I) Covariance matrix of input noise. II) Covariance matrix of measurement noise. III) Hyper-parameter of UKF. For the parameter learning. one of discriminative training methods is adopted to obtain the optimal parameters automatically. Simulation and experimental results in an outdoor environment are presented. We demonstrate the effectiveness of our proposed learning method for localization.

researchmap

Event date： 2010

researchmap

Event date： 2010

researchmap

Event date： 2010

researchmap

Event date： 2010

researchmap

Event date： 2010

researchmap