To develop a small-size 3D sonar for Autonomous Underwater Vehicles, a test system, which consists by a T-type acoustic transducer array and signal processing units, is constructed. The signal processing units are controlled by a laptop PC, and can generate a special waveform to transmit, and record multi channel signals for post processing simultaneously. In this paper the design of the acoustic transducer array and the results of fundamental experiments in a test tank are described.

抄録全体を表示

Autonomous Underwater Vehicles are valuable platform for the oceanographic survey. To develop the software of AUVs, it is important to use testbed vehicle to perform experimental development, because the computer simulation is not reliable to the influence of complex flows around the vehicle and unknown obstacles. This paper describes the design of a small testbed AUV "KOLABOT" for software development. By using plastic materials the vehicle is light weight and easily reconfigured. Results of the Zigzag tests in a test tank have validated the system performance.

抄録全体を表示

In general, primary inspection work for under side of piled pier floor slab is executed by divers or workers with stiff. Recently, ROVs also are often used for the primary inspection work. However, the vehicle operator can not easily grasp the vehicle operation status such as vehicle motion and position in the environments of intricate constructions and overhead enclosed, and this problem often impedes smart operation. In this research, I propose a teleoperation assistance system which estimates and indicates the vehicle motion by using captured continuous ceiling images. In this paper, tank test result using experimental system of the assistance system is described.

抄録全体を表示

This paper describes a dual arm underwater robot system. To solve energy/environmental problems or to effectively use of marine resources, various robotic operations are strongly required. In particular, dual arm operations are the basic requirement. We have developed a ROV (remotely operated vehicle), which has sixteen degrees of freedom including a dual arm system with 10-DOF. Some handling operations are experimentally demonstrated.

抄録全体を表示

The final goal of this study is to develop a buoyancy control device for underwater robot utilizing phase change of material. A conventional underwater robot and a conventional submarine adjust buoyancy by discarding a set of iron balls or seawater in other place, which is harmful to the environment. Therefore, we proposed new buoyancy control method utilizing phase change of material. In the previous papers, we used paraffin wax. However, its melting point is high (approximately 47℃). Therefore, we must look for a new material that has lower melting point, and not harmful to the environment. We manufactured a new buoyancy mechanism to test materials. Finally, we focused on lard and find out that lard has lower melting point (approximately 25℃) with high density difference between liquid and solid phases and thus response faster than paraffin wax.

抄録全体を表示

The goal of this study is to develop a buoyancy control device for underwater robots based on the hypothesis on the buoyancy control mechanism of sperm whales. This mechanism utilizes volume change caused by phase changes of materials. In this paper, we report a new piston type buoyancy control device using a two-step piston. Paraffin wax filled in the device was cooled by pertier elements and heated by them and a nichrome wire, resulting in volume change. We conducted experiments and confirmed that the proposed mechanism could cause the amount of buoyancy change.

抄録全体を表示

The propulsion method based on the aquatic life has many advantages as compared with the screw propeller. We have been developed Eel-like swimming robots of prototype 1 and prototype 2. The efficiency of swimming behavior depends on the shape of the robot in the water. In this paper we report the effect of changing the head shape, entire length of Eel-like swimming robot, and the number of the progressive wave. The miniaturized head reduced fluid resistance and improved swimming speed of an Eel-like robot. Prototype 2 robot swims underwater completely and on the surface of the water. In the case of swimming by serpentine movement like snakes and eels, furthermore, the swimming efficiency is affected by the shape of the cross section of the body. So we changed it into circular and rectangular cross section, and investigated the effect to the swimming form and the swimming speed. The number of the progressive wave has as big influence on the swimming performances as robot shape.

抄録全体を表示

Thrusters that have a single propeller are used for underwater robots. The thrust force of the thruster differs according to the rotational direction of the propeller and, since the thrusters for underwater robot generally use magnet couplings, the rotational speed of the propeller cannot be measured directly. Then we are developing a prototype thruster with front and rear propellers. The thruster has sensing parts of rotational speed of propellers. In this paper, we present our developing thruster and show the basic experimental results of the thruster.

抄録全体を表示

An underwater glider with independently controllable main wings for applications in oceanographic surveying, monitoring of the marine environment and ocean exploration is under development. Power-driven propulsive machinery is unnecessary for this underwater glider so that it is applicable to long-term and wide-range operations. The purpose of this research is to offer an underwater glider with high manoeuvrability and this paper deals with motion simulation of an experimental vehicle named 'ALEX' as a test-bed for an underwater glider with independently controllable main wings. ALEX was designed to realise high performance manoeuvrability. Results of the motion simulation were compared with the experiment. The authors have concluded that a feasible model could be verified and an 'underwater glider with independently controllable main wings' will show high performance of motion.

抄録全体を表示

In this project a program that is capable of generating one image of an entire water pipe's inside wall was developed. An Autonomous Underwater Vehicle (AUV) equipped with a fish-eye camera will be deployed to dive through the pipe and take pictures of its inside wall. A computer program was implemented that maps such pictures to the 3 dimensional model of the water pipe and based on that, generates the image that one would see if the photographed section of pipe was cut along the side and rolled out. A second algorithm uses the rolled out versions of a series of consecutively taken photos to generate a mosaic showing the entire pipe in a single picture.

抄録全体を表示

In this study, drive system for micro-robot using IPMC(Ionic Polymer Metal Composite) and energy supply method using external magnetic field are approached. Last fiscal year, driving test for thrust measurement of fish robot with the IPMC fin actuator and the drive circuit design for energy supply by external magnetizing field had been done. Based on last fiscal year's research, performance assessment of the drive IC and the drive circuit were done in this fiscal year. In addition, to check whether to obtain assumed voltage and current, performance assessment of magnetic field impression device was done. As a result, the problem of the drive circuit and the improvement point of the magnetic field impression device became clear.

抄録全体を表示

Underwater caves are paid attention in the field of the natural history science and the hydro geomorphology. However, the investigation is not advanced to a lot of underwater caves because it is dangerous. We developed the Untethered Remotely Operated Vehicle (UROV) "KAMANEKO" that is specialized for underwater caves exploration. We produced a compact UROV to use it in a narrow area. The UROV has only two thrusters for the horizontal migration, it inclines by the gravity moving system, and it submerges and surfaces. We challenged the investigation that used UROV for the first time in the Ojika-Do cave, and newly discovered unknown surface, stalagmite and limestone pillar. It was understood that this UROV method is effective and efficient.

抄録全体を表示

Impact of climate change causes serious problem such as drastic decrease of marine resources even in shallow water area. Compact Underwater Vehicle is suitable for their observation since it need not much manpower. We propose a new type of Compact Underwater Vehicle called Hybrid-type Underwater Vehicle, which has the feature of both Remotely Operated Vehicles (ROV) and Autonomous Underwater Vehicles (AUV). To verify the performance, tank experiments were carried out. The results have shown good characteristics.

抄録全体を表示

Recently, it is claimed that Earth environment is getting worse and worse all over the world and global heating also gives huge influence to the ocean. IPCC (Intergovernmental Panel on Climate Change) forecasts a decrease in shelf of coral, a change of ecosystem by the acidification of the ocean. Searching the ocean is very important but difficult thing because underwater is extreme environment for human beings. Underwater vehicles which take underwater operation for human being have been researched and developed by researchers worldwide. Because they have demands of the military industry, the oil industry, and so on. On the other hand, a Japanese market of underwater vehicles is smaller than European and American one. Then, we designed and developed an ROV (Remotely Operated Vehicle) for Japanese market, and evaluated low level controller of the ROV.

抄録全体を表示

In this paper, we developed a self-localization algorithm for Autonomous Underwater Vehicles (AUVs). AUVs have six DOF and can move freely without constraint of any cables such as power lines, control lines, and signal lines. Thus, AUVs need high accurate self-localization on dead reckoning navigation using an Inertia Navigation System (INS) or an acoustic sonar system. However, high accurate INS is expensive and acoustic sonar system needs a lot of efforts to recognize self position through real-time measurement. Therefore, we aim to develop a real-time self-localization system which is reasonable and robust using visual information from monocular camera.

抄録全体を表示

In this paper, we propose the X4-AUV which is a new model of autonomous underwater vehicles. This model is controlled by only four thrusters. Since the X4-AUV has underactuated and nonholonomic features, a kinematics based control law is first transformed into a chained form. Next, the transformed kinematic model is controlled by using a manifold. A backstepping method that is one of adaptive control methods based on Lyapunov methods combines the kinematics model and the dynamical model to construct the torque control of X4-AUV. Finally, computer simulations are shown to demonstrate the effectiveness of our approach.

抄録全体を表示

Underwater Vehicle-Manipulator Systems(UVMS) are necessary for ocean exploration. Many studies have been done about UVMS, however the experimental studies are only a few and the control methods are based on Autonomous Underwater Vehicles. We have proposed continuous and discrete time Resolved Acceleration Control (RAC) methods for UVMS. The effectiveness of the proposed methods have been shown by simulations and experiments using a floating underwater robot with 2 dimensional and vertical planar 2-link manipulator. Furthermore, we have also proposed RAC methods with disturbance observer for robot base. In this paper, experiments of the 2-link underwater robot using a discrete time RAC with disturbance observer for robot base and manipulator are done. The experimental results show the effectiveness of the control method.

抄録全体を表示

Underwater Vehicle-Manipulator Systems (UVMS) are expected to make important roles in ocean exploration. It is considered that UVMSs will be operated by automatic and manual control. We have been proposed the automatic control method. In this paper, we propose a master-slave system for UVMS.

抄録全体を表示

As the tools for ocean survey and natural energy resources search, underwater robots are expected to operate these missions because underwater operations are too danger for humans due to water pressures. In fact, Autonomous Underwater Vehicle (AUV) succeeded some operations such as observations of an underwater volcano and a Sperm Whale tracking. One of the key issues for underwater robots is self-localization problem in the ocean. Self-localization systems are required to navigate underwater robots. We focus on Acoustic Positioning System because it uses sound waves which propagate well underwater. However, any APS has problems caused by a kind of noise "multipath" in general. Accordingly, robust APS against noises is hoped for underwater robots. In our previous research, we described a 2-dimensional underwater sound source localization system designed for mounting in underwater robots. In this paper, we mention about an expansion to a 3-dimensional sound source localization.

抄録全体を表示

We have developed easy transfer system between floor to floor of cleaning robot. The characteristic of this system is not necessary for reorganization of elevator. This system is composed by "communication" equipment, "part of actuator." Each of the equipments is communicate with information of destination floor from robot. And these are remote controlled by pushing a button in the elevator.

抄録全体を表示

The purpose of this research is to develop an on-line person following method for a mobile robot. The robot generates a robot-centered local map describing dynamic and static objects and plans a safe path. We set the position of the person to follow as a goal position, and other objects are considered obstacles. The skeleton of the current free space in the local map is calculated, and then a potential field is calculated such that the potential increases towards the goal position along the skeleton. Finally, a path is generated by a kinodynamic motion planner which uses the potential with consideration of the kinematic and the dynamic constraints of the robot. Since dynamic objects continuously move, a path is replanned as new sensor data arrive. We use a laser range finder and a stereo camera for detecting static and dynamic objects, respectively, in the experiments. We have succeeded in making the robot follow a target person among various dynamic environments.

抄録全体を表示

すでに我々が開発した,清掃ロボットをFig.1に示し,ロボットの主仕様をTable1に示す.本ロボットは,主にオフィスビル等の高層ビルの床面清掃を行い,エレベータを乗り継ぎながら複数階の清掃を行うシステムとなっている.オフィスビル等のカーペット床では,細かい砂や埃といったゴミが多く,これらのゴミを吸引清掃しながら移動していく.ロボットの車体寸法として,エレベータ乗降を自動的に行うため,エレベータの開口部幅800[mm]程度の狭い場所でも通過な大きさとなっている.

抄録全体を表示

In this paper, we describe a robot system to control human detection. In our proposed system, the robot surveils a security area with equipped SOKUIKI sensor, ofdinarily. When it detect trespasser who is for the area, robot calculates his velocity vector, plans own path to forestall and interrupt him and starts to move along the obtained path. While moving the robot continues these processes for adapting change of situation. After arriving at a trespasser's opposite position, the robot advises him not to be headed more and change his course. Our experimental results show the robot can estimate human's future trajectory and calculate appropriate its own path.

抄録全体を表示

In this paper we present a method of self localization in the RoboCup middle size league based on the field lines. This method estimates position of the robot by using Monte Carlo Localization (MCL). MCL is widely used for mobile robots localization in a known environment due to its good real-time performance and robustness. MCL estimates state of a robot based on sensory data. In our algorithm, the sensory data are distances to field lines and wheels odometry measurements. Usually MCL estimates robot state (x,y,θ). We present dimension reduction method of state with simple direction estimation algorithm.

抄録全体を表示

On the RoboCup Middle Size League, official rules have been changed every year to aim at the task that playing with human. For instance, color landmarks such as Balls or Goals are disappeared or the playing field becomes wider etc. Corresponding to this, the techniques other than the past method which uses omni camera and color information are required. Then, in this paper, we describe about an object detection technique that uses the stereo camera and the edge based model named "cylinder model". By this technique, it became possible to detect the field player far from a camera by comparison past method, without color information.

抄録全体を表示

In this paper the design method of new omni-directional mirror by using progression is proposed. As compared with the picture obtained by hyperbolic omni-mirror, the picture with better sharpness of the field was acquired in wider area. In addition, the shape of a mirror can be realized using distribution of arbitrary resolution, because the distance that reflects to the selected pixel can be chosen freely.

抄録全体を表示

This paper presents a novel Self-Localization method for RoboCup soccer robot. For the robots, self-localization methods with using white lines have been proposed and generally used. However, these methods have a weakness for lack of white line information and have possibility that the robot cannot calculate an accurate position of the goal because of a mistakenly calculated self location. In this paper, a self localization method based on goal posts recognition is proposed. The robot can calculate their position on the absolute coordinate independently of precision of white line detection. By means of accurately detecting of goal posts, the robot can move normally without specific color of the goal. In this method, there are some steps of conditionality about distance and angle. To verify the effectiveness of the proposed method, some experiments including match at RoboCup MSL using a GK robot were carried out.

抄録全体を表示

In this research, we developed a system which determines the team behavior according to environment. When the environment which surrounds an agent changes dynamically, it is difficult to determine individual behavior supposing the strategy and task on all the conditions. Therefore, it is necessary that agents guess and decide the behavior from the past experience. According to the conventional system, we develop the multi-agent system which is based on Self-Organizing Relationship (SOR) network. And also the learning data is made by our experiment. In order to verify the usefulness in the real environment of the proposal system, the behavior determination system was experienced in the soccer robot for RoboCup Soccer Middle Size League.

抄録全体を表示

In recent years, the delivery support robots are expected in coexisting environment with human being. Self-localization of mobile robot is necessary in order for robot to perform task autonomously. Many researches about robot localization have been done using many kinds of sensors, thus increased the cost of the system. How to lower the cost of the robot system is the main topic in order to promote robot applications from a practical point of view. In this research, we propose the robust method of self-localization for mobile robot using cheap USB camera to recognize the landmark in the environment. The proposed method uses SURF that is the robust method to recognize landmark, then positions the mobile robot based on the results of SURF.

抄録全体を表示

Recently, cleaning robots for domestic use are developed. However, present algorithm cannot always sweep the workspace uniformly, and this causes the inefficiency of the algorithm. The best sweeping algorithm may depend on the shape of the room. To select the best algorithm, we examined a method to acquire the shape of a room in the last report. In this report, we adopt an actual robot model so that the robot can distinguish rooms in the real world.

抄録全体を表示

This paper describes the localization of a guidance robot and humans to realize the robotic transportation system for shopping assist. Cameras set at the ceiling to cover the corridors of the shopping store are used as sensors. Each camera image is processed to detect not only the robot but also humans. Detected positions of the robot and humans are sent to the environmental database through LAN. The experimental results are reported using the model shopping store to show the performance of the method.

抄録全体を表示

The robotic transportation system for shopping assist has been developed. The system consists of a guidance robot, a cart robot, and environmental cameras. The guidance robot is an autonomous mobile robot with position localization function, the cart robot with a shopping basket can follow the guidance robot, and environmental cameras can detect the robots and human. Verification test has been performed in the test environment assumed for a food store. The guidance robot was able to guide a person to his/her desired place and the cart robot was able to carry the foods while following the person.

抄録全体を表示

We developed a simulator for robocup mid-size league that can draw omnidirectoal vision for each robot. By using OpenGL's cube map texture, environmental image around robot can be reflected on a sphere mirror. We also developed an intelligent part and image processing part using omnidirectoal image in this simulator. Various image features such as another robot figures and line on a field can be used on the development of our program.

抄録全体を表示

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

抄録全体を表示

In RoboCup middle-size league (MSL), we have been using the positions of color goals and corner poles as landmarks, and localized the position of mobile robots. According to the change of the regulation of RoboCup MSL, we obliged to use other objects as landmarks since color information has been removed. This paper describes the self-localization method based on position of lamps on the ceiling as the landmarks. Provided that the ceiling is horizontal and the lamps on the ceiling are allocated along a straight line, the position of a mobile robot can be specified. We conducted experiments in the corridor and confirmed the validity of the idea.

抄録全体を表示

We participate RoboCup humanoid league and middle-size league. Color information is specified in RoboCup regulations in order to ease object recognition, yet color recognition is one of the big issues because illumination condition varies depending on the competition site. We examined a simple yet robust color recognition technique against illumination condition change. The position of specified color area is once taught by human operator, and the corresponding cubic region in HSV space is used as a color template and is used for color recognition. Depending on the change of illumination change, the different region in HSV space of the specified area is used for color matching. We conducted color recognition experiments in both indoor and outdoor environment.

抄録全体を表示

This paper presents a method that utilizes state value functions of macro actions to explore appropriate behavior efficiently in a multi-agent environment. First, the agent learns a few macro actions and the state value functions based on reinforcement learning beforehand. Second, an appropriate initial controller for learning cooperative behavior is generated based on the state value functions. The initial controller utilizes the state values of the macro actions so that the learner tends to select a good macro action. By combination of the ideas and a two-layer hierarchical system, the proposed method shows better performance during the learning than conventional methods. This paper shows a case study of 4 (defense team) on 5 (offense team) game task, and the learning agent (a passer of the offense team) successfully acquired the teamwork plays (pass and shoot) within shorter learning time.

抄録全体を表示

We developed an autonomous target-following robot, under complicated environment which includes various shaped obstacle. Its special features are, the simple algorithm which doesn't use map data nor position identification, the deformed body shape to be long backward for taking baggage, using ultrasonic sensor as the primary sensor for reducing cost. We optimized the ultrasonic sensor positions and improved trajectory generation algorithm; therefore, it was achieved that the robot implements target-following with obstacle avoidance under the limited complicated environment. Moreover, it was confirmed by experiment.

抄録全体を表示

This paper proposes a sound identification method for a mobile robot in home and office environment. We propose a simple sound database called Pitch-Cluster-Maps(PCMs) based on Vector Quantization approach. Binarized frequency spectrum is used for PCMs codebook generation. It can describe a variety of sound sources, not only voice, from short term sound input. The proposed PCMs sound identification requires several tens(msec) of sound input, and is suitable for a mobile robot application which condition is dynamically changing. We implemented the proposed method on our mobile robot audition system equipped with a 32ch microphone array. Robot noise reduction using proposed PCMs recognition is applied to each input signal of a microphone array. The performance of daily sound recognition for separated sound sources from robot in motion is evaluated.

抄録全体を表示

The care system for elderly people is expetcted to be implemented in aged society. The system needs both of tracking a human and observing him/her closely at the same time. We have develeped the robot head with zoom cameras. We combined the head and wide angle cameras in order to enable both of tracking and observing at the same time. It can track a human by using wide angle cameras and the result of opticalflow calculated by the image processing module. Then we calculate position of a human by stereo matching and move zoom cameras towards him/her and control the zoom factor according to the distance to him/her. In this way, we enable both of tracking a human and observing him/her closely at the same time.

抄録全体を表示

Recognition of objects by robot have two problems. First problem is that the method of registering objects is not easy in general. Second problem is that it is difficlut for normal cameras to recognize distant objects. So we used the robot head, developed in our laboratory, with the active vision system with both of wide vision and high visual auity in order to resolve these problem. It has two active zoom cameras and two stereo wide angle cameras. We can register objects by showing the robot objects and clipping images twice with the information of distance and SIFT features. And the robot can recognize variable objects in the room by moving zoom cameras, changing zoom factor and using SIFT features.

抄録全体を表示

This paper describes a vision-based interface for instructing robot motion easily. An interface that makes the robot move in the same way as user's motion is effective for an intuitive motion instruction. Such an interface can be realized by estimating the pose (position and orientation) of the interface and executing move commands for making the robot take the same pose. We estimate the pose of the interface by a monocular SLAM method, which is based on visual features and the extended Kalman filter. By additionally using an orientation sensor and an accelerometer, the reliability and the accuracy of pose estimation are improved. From the estimated pose, the target values of the robot position and the head orientation are set and the robot moves to achieve them. We implemented an experimental system which run in real-time (30[Hz]) and successfully applied it to controlling a humanoid robot.

抄録全体を表示

We propose new input method of operation. Instructions to robot are given by stepping on a part of operation screen that is projected on running surface by projector equipped on the robot. This paper examines the layout of instruction buttons on operation screen, and the discrimination of obstacles from stepped foot to apply autonomous movement and the path search for obstacle avoidance are also discussed.

抄録全体を表示

We propose new input method of operation. Instructions to robot are given by stepping on a part of operation screen that is projected on running surface by projector equipped on the robot. This time, we designed the human-friendly amusing mobile (HFAM) function that is new usage of the robot equipped with the SOI. There is a scenario that the robot and a child will be able to play together. In this paper, the outline of the HFAM function and the technology are discussed.

抄録全体を表示

In automobile production lines, there are many operations by not only autonomous machinery as industrial robots but also man-machine cooperative systems as power-assist systems. In this case, efficient mode switching from human-oriented operation to machine-oriented operation is important. This paper presents a new mode switching control method from the manual mode to the automatic mode for flexible parts conveyance. In the manual mode, impedance control is applied to the power assist devise. In the automatic mode, the PID control and the desired trajectory of the power assist machine are used to suppress vibration of a flexible part. The desired trajectory is calculated by using minimum crackle trajectory. The minimum crackle trajectory is an ideal trajectory in reaching task of a flexible object by human hand. The effectiveness of the proposed method is verified experimentally.

抄録全体を表示

We have proposed an assist control method for flexible parts conveyance using a power assist machine. The proposed assist control is based on variable impedance control. The impedance parameters are identified by applying recursive least square method to the experimental data with good performance. In this paper we investigate the effectiveness of proposed method by analyzing vibration of a flexible part on the strain - strain rate phase plane. Moreover, the effect of inertial force of the power assist machine to the flexible part is analyzed on the phase plane. As a result, the effectiveness of the proposed assist control is verified for flexible parts conveyance.

抄録全体を表示

The purpose of this study is the development of the force-and-displacement detection interface that can contact with environment without the contact stability problem. The proposed interface is set at the tip of the robot arm to detect the control force applied by the human operator and the deflection between the tip of the robot and the human operator. The proposed impedance control scheme of the robot motion can contact stably with the external environment under the impact force. To confirm the effectiveness of the proposed control system, a human-robot cooperative contact experiment is carried out. Experimental results show that the proposed impedance control scheme can provide the adequate motion against the impact force without the contact stability problem of the impedance controller.

抄録全体を表示

For directly aiding human life, we develop a small six-axis robot arm that is easy to incorporate in daily life. We suggest a task teaching method in which we move the robot arm directly by hand and thus can control it more easily. To reduce the quantity of data for the task teaching, we suggest the method of taking out each angle data at the time intervals that are obtained from angular acceleration. From simulation, the Hermite interpolation is appropriate for the suggested method. By using this method, it became possible to calculate the angle instruction value with high reproducibility and a small number of data.

抄録全体を表示

In this paper, we propose Multijoint Control Reduction Algorithm Model for multijoint robot. The method reduces inverse kinematical operation. Furthermore, this method makes possible the construction of an easy control system, because the use of the Fuzzy Logic enables the linguistic modeling of the joint angle. In the Experiment of horizontal coplanar points with an arm of small humanoid robot, we confirmed that a robot performed the same movement with a person. In addition, we realized fast information performance in control algorithm of the all integer arithmetic to the low power microprocessor.

抄録全体を表示

In this study, we analyzed a pedestrian's walking pattern and designed an avoidance control method for a robot moving among pedestrians. By the proposed method, a pedestrian's hidden foot movement was estimated by the extended Kalman filter, and the robot escaped collision with the pedestrian by using the proposed voluntary guidance control method with the estimated values. The effectiveness of the proposed method is shown through experiments.

抄録全体を表示