Walking abilities tend to be declined by getting old. The decline of walking ability causes serious problems such as falling down and breaking a bone. Moreover, it sometimes becomes one of the essential factors of death. Therefore, in this paper, we propose a system that estimates their walking abilities by using Kinect, which is an infrared depth sensor for tracking skeleton model of a user. Based on the information of skeleton model, the system analyzes elemental indices such as walking speed and step length. Then it estimates a decline of walking ability, which is used for detecting the signs of falling in the early stages.

View full abstract

In order to improve physical function of elderly people, they need to get physical training. There are activities which require integrated lower body and arm. However there are no quantitative testing methods of the degree of recovery for the coordination between arm movements and lower body movements. In this study, we focus on dart throwing motion to suggest the quantitative evaluation of the coordination between arm movements and lower body movements in dart throwing motion. Joint coordination was analyzed with uncontrolled manifold (UCM) analysis at different throwing distances. It is suggested that the joint coordination in order to achieve the same elbow position is mostly required in dart throwing motion.

View full abstract

A human foot complex consists of a flexible flesh and many muscles, ligaments, and bones. Structured compliance provided by the complex structure is supposed to contribute to efficient and stable bipedal locomotion. Detailed behavior of the foot complex during walking, however, is not well understood . We built a gait simulator driven by pneumatic musculoskeletal system in order to clarify the mechanism of a human walking by using the gait simulator with a cadaver foot. This gait simulator duplicates the double support phase in experiment environment. We also made realistic foot model with bones and artificial skin as a test model of new robot foot for testing the performance of the gait simulator.

View full abstract

Recently, the rubber hand illusion attracts much attention in order to examine human brain mechanism of the sense of ownership. However, most previous research have only focused on the conditions for the occurrence of the illusion. In this study, we propose a new method to measure the occurrence time of illusion.

View full abstract

A human running motion can be modeled by a spring-loaded inverted pendulum (SLIP). However, the SLIP model does not include human-like pelvic movement which can contribute to increase a jumping force and absorb a landing impact. Based on a running gait analysis, we propose a novel model, named SLIP2 (Spring-Loaded Inverted Pendulum using Pelvis), that is composed of a body mass, a pelvis and leg springs. This model can store elastic energy during running by utilizing pelvic movement in the frontal plane. Additionally, we developed pelvic oscillation control and running velocity control. To evaluate the proposed model and control methods, we performed hopping and running simulations. The simulation results show that the SLIP2 model achieved hopping and running motions. Moreover, it shows that the difference of the pelvic oscillation phase affects the jumping force.

View full abstract

In general, the most difficult aspect of realization of cooperative behaviors between robots, as well as between a robot and human, stands on how one acquires another dynamical, mechanical, and control model. For learning the other models, it is necessary to dissociate whether the resulting motion are by self or other operations, that is to say, self-other distinction in learning. To discuss the mechanisms of self-other distinction, we consider the cooperative manipulation task by two arm-robots, in which robots are physically interact with each others, and are equipped for independent controllers. In the simulation the feedback-error learning is used for acquisition of required motions. In this study we show that the minimization of various parameters concerned with the controlling arms, along with typical learning, can lead the feedforward controller which is nearly independent of the other motions.

View full abstract

Humans generate adaptive locomotion according to locomotion speed, physical characteristics, and environment. This ability is achieved via the coordination of the redundant degrees of freedom in their body. For generating such adaptive locomotion, postural stabilization through synergetic activation of muscles is a fundamental function. In this study, we propose a novel posture control based on synergetic activation of muscles. The simulation results show that standing posture stabilization is achieved by reduced neural commands.

View full abstract

In contrast to the high adaptivity of quadruped animals in many environmental conditions, it is hard for conventional quadruped robots to behave in complex environment conditions. In this paper, for the purpose of investigating the adaptivity possessed by animals' musculo-skeletal system, we built a bio-inspired quadruped robot named "Pneupard" that duplicates a feline musculo-skeletal system of hindlimbs and has 16 pneumatic muscles and 6 DOF. We proposed a sole reflex-based control and verified its effectivity by conducting a walking experiment, in which the robot performed a stable walking with a two dimentional restriction. In addition, we conducted another experiment, finding that the robot with the proposed control could walk stably on irregular terrain.

View full abstract

We clarify the design principle, i.e., relationship between terrain condition and locomotion style of a robot, using an evolutionary robotic system that designs both morphology and intelligence simultaneously. Here, rough terrain consisting of many obstacles is defined by the height and the arrangement. The robot designs own morphology and control system, that is, locomotion style, to overcome such rough terrain by a genetic programming. The simulation results showed the following design principles: (1) When the height of the obstacle was high, the arrangement noise of the obstacle dominated the emergent mechanism of the wheel and creeping locomotion styles. (2) When the height of the obstacle was low, the wheel locomotion emerged regardless of the arrangement noise of the obstacle.

View full abstract

Recently we developed a non-invasive braincomputer interface (BCI) to allow an user to control an humanoid robot within a complex scenario involving locomotion and object handling. The main challenge lies in the limitations of such BCI: limited output commands, low bandwidth and possible interpretation errors. Thus we developed several techniques to overcome these limitations by taking advantage of the embedded autonomy of the robot and putting it to use within the user interface. In this paper we present two of these techniques. The first shows how the vision system can be used to allow the user to grasp objects using the robot; a task otherwise extremely difficult using non-invasive BCI. The second illustrates how the vision system can also be used to ease the steering of the robot. These techniques illustrate the way we aim to develop better solutions for humanoid control using a brain-computer interface.

View full abstract

An optical system which can control its gaze faster than a previous method of actuating a camera itself has been rearized by actuating mirrors in front of the camera. If we use the optical system for measurement, we need to know precisely optical parameters such as a optical center of the camera, but it is generally difficult to measure them exactly. In this paper, we define the optical parameters of this optical system, and propose a bundle adjustment method in this optical system to estimate the parameters. Defining projection matrices in several different designated gaze directions only by the optical parameters, a nonlinear minimization problem of reprojection error in this bundle adjustment becomes smaller than treating postures and positions in all images as unknown, and we can get the interesting parameters explicitly without decomposing the optimized projection matrices. The experimental results in simulation show that almost all parameters converge to true values nearer than initial values.

View full abstract

In this study, we proposed a motion detection circuit based on the biological vision system. The proposed circuit is constructed with the analog part and digital part. Analog circuits generate edge signal. Digital circuits generate motion signal. The simulation result with the simulation program with integrated circuit emphasis (SPICE) and the proposed circuit is a simple structure. The measured results of a fabricated circuit showed that the proposed circuit can output the pulsed voltage (motion signal) when an object moves on the circuit. In the future, can be expected to achieve new compact motion detection sensor by using the propose circuit.

View full abstract

SIFT matching is a quite robust matching approach for the usage of robust SIFT descriptor. In order to overcome the disadvantages brought by gray images, many color SIFT methods have been proposed. However those methods mainly focused on the improvement of SIFT descriptors. The excessive number of feature points, especially those useless points on background rather than a target, on the other hand, is not paid enough attention yet. This paper proposes an approach to improve the efficiency in SIFT matching. Using color filter to eliminate feature points with different color from those on a target. Therefore much calculation on descriptors and matching could be reduced so that the efficiency could be improved.

View full abstract

Optics, such as a camera lens, works as the first filters for vision systems. It projects 3D light sources distribution on a planar imager. To optimize image processing performance, adaptive control of optics must be essential. However, most conventional vision systems use fixed imaging optics. This is mainly due to the slow response of the imaging optics. In this presentation, a new vision system, named "Roboptic Vision", based on a coupling of high-speed optical devices the authors have developed and high-speed image processing is proposed. Totally millisecond-order response of whole system enabled various applications such as 1ms auto pan/tilt, arbitrary focused image synthesis and 3D motion estimation from focal depth scan image sets.

View full abstract

It is very important for robot to find objects, especially for a home service robot which works in human living environment. To find the environment contains many kinds of objects, we need the object recognition method using large dataset. Random Forest, developed by Leo Breiman[1], is an ensemble classifier composed of many decision trees . The method outputs the class that is the mode of the classes output by individual trees. Each trees can be calculated independently. So, the algorithm runs efficiently on large databases. Thousands of input variables can be handled without variable deletion. In this paper, we present a method for the object recognition by random forest using a large dataset of RGB-D images. And we evaluate its acuuracy and computationnal cost.

View full abstract

In this study, we developed a 1000-fps stereo laser tracking system that can simultaneously track a fast moving object and draw an arbitrary laser pattern on the surface of the tracked object. Our system can always control the direction of laser beam with a 2-DOF galvano-mirror along a given trajectory on the surface of the tracked object by processing stereo 512X512 images in real time at 1000 fps. The performance was demonstrated by showing an experimental result when a triangle pattern was laser-drawn on the surface of the tracked object moving back-and-forth in three dimensional space.

View full abstract

This paper describes a method of recognition for transparent objects on the desk, using a monocular camera. First, we apply a machine leaning-based method using edge patterns to the transparent desktop object recognition. Edge patterns which appear in the local area of the image are represented by a probabilistic topic model. Edge pattern features indicating presences of specific edge patterns, are used for training a Support Vector Machine (SVM) binary classifier, which dscriminates between transparent objects and backgrounds. Then, we try to detect transparent objects in desk scenes using the classifier. Next, we define a highlight pattern feature describing highlight regions of objects. Highlight pattern features are represented by binarization of highlighted pixel ratio in local areas, and combined with edge pattern features to enhance the previous method. Finally, combined features are tested in cross-validation and object detection and shown to be superior to the previous edge pattern features.

View full abstract

This paper presents a new categorization method for typical indoor places (e.g., "office", "kitchen", "corridor") using 3D scans by a laser range finder. The proposed technique utilizes panoramic range and reflectance images obtained by a laser range finder. Firstly, the panoramic range and reflectance images are divided into several partial images, and histograms of local binary patterns (LBP) are created from each partial image. The place category of each partial image is determined by support vector machines using a feature vector which concatenates the histograms. Finally, the place category for a panoramic image is determined by voting the decision of partial images. Experimental results show that the recognition rate of the place categorization is improved by the proposed method using panoramic images compared with the method using partial images.

View full abstract

This paper describes a real-time background subtraction method for high-speed object tracking with a millisecondorder auto pan-tilt system called "Saccade Mirror", where background changes according to the pan-tilt angle of gaze. This method relates each background point to spherical coordinates so that background information can be easily registered and reconstructed. We also propose a way to deal with motion blurred background images. Several experiments were conducted to validate our proposed method.

View full abstract

In this paper, we propose a robot manipulation system controlled by a stereo high-speed vision system. The vision system recognizes targets by using CAMshift. CAMshift is an tracking algorithm based on the pattern of the color histogram of a target, and it is robust against disturbance of image noise. In the experiment, the system recognized the positions of both a hand of a human operator and a stick grasped by the hand, and estimates the position and orientation of the stick. The robot arm was controlled so as to follow the motion of the stick. The motion of the arm was quick and responsive. The proposed system is useful for human-robot interaction systems.

View full abstract

Focal length measurement mechanism is essential for tunable liquid lens to realize new functions of vision system, such as high-speed zooming system. In this paper, a preliminary focal length measurement mechanism is reported.In the mechanism, laser displacement meter was adopted as a measurement component so that high-speed response time and high resolution could be achieved. Through experiment, result that suggestion mechanism worked depending on focal length change was confirmed.

View full abstract

Separating specular reflection components from diffuse reflection components is an important subject in computer vision. Hence, the author's group has developed a method to separate a diffuse image from a high-speed-video image with specular reflection components. The method uses an estimation technique based on luminance variation owing to flicker of a strobe. This study extends the abovementioned method to a real-time system. Although this method needs a high speed camera with a short sampling time, it is possible to show a specular reflection free image in real-time since the algorithm requires small calculation amount.

View full abstract

Interventional Radiology (IVR) is a kind of surgical operation using real-time CT images and needles. IVR has an advantage that it is a lower invasive operation than a general surgical operation. On the other hand, IVR has a disadvantage that operators are exposed to radiation of X rays from a CT machine while they conduct the operation. To solve this problem, we propose developing a serial link robot as a remote controlled robot. In this paper, prototypes of robotic interventional radiology are described. First, specifications for the robotic IVR are described. Next, first version of prototype is designed and results of preliminary experiment for artifact are shown. In the second version, we have constructed a 3DOF mechanism and its driving system which is commanded by a joystick controller. Finally, experimental results of the developed prototype system are shown.

View full abstract

Locally operated detachable end-effector manipulator (LODEM) that can support single port endoscopic surgery when operated by doctors working near the patient was developed. The 3rd arm with commercial forceps for handling internal organs collaborates and coexists with the surgeon in the clean area. We proposed a new mechanism of differential driven ball-screw for the change of orientation angle of the commercial forceps. It had four axes: pitch, yaw, roll, and open/close. The mechanism was mounted on passive gimbals at the tip of SCARA typed LODEM and driven by wires and cable-rods. We made a prototype and the dynamic characteristics were experimentally measured. The position accuracy of pitch or yaw axis was 0.4 mm respectively. The time delay of roll or open/close axis was 0.4 s respectively.

View full abstract

Laparoscopic surgery has been used in various surgeries because it brings many benefits for patients such as smaller scar, less pain, and shorter hospitalization compared with open surgery. Moreover recently, single-incision laparoscopic surgery (SILS), in which surgical instruments are all inserted into a single incision, provides more cosmetic benefits than laparoscopic surgery although it has problems like the narrowness of the operating range, the interference between the surgical instruments, and the visual obstruction of the surgical site. In order to overcome the difficulty of SILS, we proposed a "patient-mounted" master-slave robot system which brings surgeons the operability similar to normal laparoscopic surgery. We confirmed the feasibility of our proposed system by making an initial prototype of "patient-mounted" master consoles and "patient-inserted" slave robots.

View full abstract

Generally, the operation of endoscope requires an experience and skill for operators, because of difficulties of correctly focusing on target region. To support their work, vision system, which can produce images to be easy to understand, is required. In this paper, we propose the all-in-focus image creation algorithm. Through the experiment, proposed algorithm could create fine all-in-focus image for endoscopic view.

View full abstract

Single-incision laparoscopic surgery (SILS) is less invasive than conventional laparoscopic surgery which requires several incisions to perform surgery. However, burden of an operator performing it increases for some reasons such as the limitation of operating ranges, interference between surgical instruments and a laparoscope and requirement of practice. To solve these problems, we developed a master-slave robot system which can conduct as if the operator was performing conventional laparoscopic surgery. We confirmed that our system would reduce the burden of surgeon in SILS by conducting simulation experiments. In this paper, in order to evaluate the proposed system with the real machine, we designed and made two kinds of flexing instruments, and we evaluated the instruments through real experiments.

View full abstract

In Neurosurgery, Opening a brain fissure allows an operator to access to basis cranii and to treat a disease part safely, but the operator is required high technique. The goal of this study is to develop a Neurosurgery simulator enabling surgeons to train opening a brain fissure procedure. Simulation of excluding soft tissue needs collision detection between a soft tissue and tools and choosing suitable contact displacement. In this study, collision detection considered its previous condition is proposed and opening a brain fissure simlation is performed with this method. In addition, the computation of collision detection is accelerated with GPU.

View full abstract

Very fine motion is required in eye surgery, more specifically, in vitreoretinal surgery, and several surgical robotic systems have been developed to assist surgeons by filtering hand tremor and improving motion resolution. However, due to human-oriented problems, the precision and safety of robotic surgery are not sufficient. In this study, we propose two methods to implement motion constraints and automatic guidance to a master-slave microsurgical robotic system. The motion constraint method limits the motion range of the surgical robot by changing the motion-scaling ratio according to the change of distance between the tip of the robotic instrument and the target. In contrast, the auto-guidance method changes the direction of the robotic motion so that the tip of the robotic instrument approaches toward the target. We conducted experiments to elucidate the feasibility of the proposed methods, and the results showed that both methods had potential to improve precision and safety.

View full abstract

Microsurgery requires elaborate and precise operation under a microscope. The master-slave microsurgical robotic system developed by our group for micro-neurosurgery successfully demonstrated the anastomosis of 0.3 mm blood vessels. However, the task completion time was longer than that of manual microsurgery because of the high motion scaling ratio (1/7) and poor usability of the master manipulators. To enhance system usability and shorten task completion time, we designed a master manipulator that can be operated without moving the operator's arms and at a low motion scaling ratio of 1/3. We performed preliminary experiments to evaluate the newly-designed master manipulator. The results suggest that the proposed system would achieve high accuracy in microsurgery while shortening the task completion time.

View full abstract

There is a surgical method called Interventional Radiology(IVR) using CT equipment in the medical treatment of lung cancer, a blood vessel, and so on. IVR has the advantage that invasiveness is low and the medical treatment in local anesthesia is possible as compared with a laparotomy surgery. Therefore the number of IVR surgery has been increasing in recent years. But, radiation exposure to doctors has been concerned in the current IVR since doctors conduct the treatment using CT equipment. Thus, the IVR robot is developed in order to avoid the radioactive exposure to doctors. In this research, a robot using a parallel link mechanism is developed from the viewpoints that hand accuracy and hand rigidity are high and a motor can be arranged in the position separated from CT scanning plane.The dual delta-type parallel link is proposed. It has the structure arranged the Delta-type parallel link mechanism with three degrees of freedom(DOF). First, the arm which has offset of the axis in consideration of interference was designed. Secondly, inverse kinematics analysis of a parallel link was conducted. Next, Delta-type parallel link mechanism with three DOF was manufactured. Finally, basic motion of parallel link robot was confirmed.

View full abstract

Myoblast cell sheets are taking into the clinical treatment of heart disorders. In our study, we proposed a surgical robot system, includes a robot system with the characteristics of Double RCM (remote center of motion) mechanism, which can cancel and compensation the beating heart surface movement for the myoblast cell sheet transplantation and two endoscopic cameras to capture heart surface motion without obstruct the transplantation of the sheets. The linear and rotation motion of the robot system are totally isolated, realized by employed a Double RCM mechanism. An experiment is conducted to evaluate the tracking accuracy of our robot system in tracking an approximate -period moving target. The tracking data is updating by Polaris system in 30Hz. From the experiment, we got the results that the linear and error is 4.67±3.73 mm and the rotation tracking error is 2.86±5.50 degree.

View full abstract

In this work, we study experimentally the effect of the non-dimensional parameters of saw-tooth surface and micro-part and the effect of amplitude vibration on the motion of micro-parts. The dimensionless quantities are functions of the range of saw-tooth pitch, p, and micro-part length, l. The experiments are performed for these scale with a range of exciting frequency applied to the surface, f. By the use of particle tracking velocimetry method, we can achieve time-dependent velocity, and then ensemble-averaged velocity of the micro-parts. In addition, the method to determine the value of amplitude vibration at a range of exciting frequency is also introduced. The experimental results show that for different l and p but the same relative scale l/p, the profiles of micro-part velocity against the characteristic surface velocity pf are similar. It seems that the motion of micro-part depends on characteristic surface velocity pf than the relative scale l/p. We also observed that the feeding velocity of the micro-part and the amplitude were maximum in the same frequency range.

View full abstract

In recent years, a large number of studies have focused on micro-robots. Moreover, yearly contests have been held in order to develop micro-robot technologies. As a result of these activities, various types of microrobot have been proposed, including motor-type, piezoelectric-element-type, and electromagnet-type micro-robots. To date, the authors' research group has been studying the locomotion mechanism and a control method for an electromagnet-type microrobot, which has a volume of 1[cm3]. The proposed microrobot has a feature of being able to turn with single electromagnet coil using the difference of the vibration characteristics of the legs. And, we have proposed a wireless remote control method for the microrobot, using the future of the locomotion mechanism and control method. In this paper, we report the results for legs' vibration experiments, which were performed to confirm the effectiveness of the wireless remote control method we have proposed previously, using a fabricated turning microrobot.

View full abstract

In this paper, we have reported that design and analysis of an in-liquid micro robot with self-energy harvesting and self-propulsion functions. There are two problems to develop micro robots with 1mm3. Those are a sticky problem and an energy source problem. To investigate solutions of those problems, we look at the natural world and found out microbes are living in liquid. They harvest energy by eating foods and solar energy and move freely in liquid; they have already solved sticky problem and energy source problem. That is our motivation to research and develop micro robot which moves in liquid. In this report, we have proposed new concept of the robot composed of voltaic cell as energy harvester and comb actuator as its source of motion. We have estimated the dynamic motion of the robot by simple linear model. We have also conduct FEM analysis to check its performance.

View full abstract

This paper demonstrates a microbial-robot that swim in low Reynolds number environments powered by flagellated bacteria. To immobilize the flagellated bacteria strongly, we utilized bacterial cellulose which is microfabricated by a new method. A large number of flagellated bacteria immobilized on the bacterial cellulose power the robotic body and a useful work is extracted.

View full abstract

An Electro-conjugate Fluid (ECF) is a kind of functional fluid, which produces a flow (ECF flow) when subjected to high DC voltage. This study develops a novel inchworm robot having sucker at its front/rear ends which are driven by the ECF flow. The robot imitates the movement pattern of inchworm, and the motion is also generated by the ECF flow. First, we developed and evaluated the sucker, which is integrated with a pressure source, and the expansion/contraction actuator. Next, we developed the robot with 68 mm long, 10 mm wide and 9.5 g weight. Finally, we conducted moving experiments of the robot on an acrylic board. The robot could move on a horizontal surface, and surfaces with inclination of 40 degrees or less.

View full abstract

This paper reports on the fabrication method of a micro solenoid coil as a receiver for MRI (Magnetic Resonance Imaging). The micro solenoid coil is consists of upside arcs and downside arcs formed by bending the flexible substrate, a selective assembly mechanism of PDMS (Poly Di Methyl Siloxane), an acrylic fixture. The downside arc of the coil was assembled selectively by fixing only the central part to the selective assembly mechanism. Since the hinge structures with low bending stiffness was located at both ends of downside arc, the cross section of the micro solenoid coil became circular. We fabricated a micro solenoid coil of 20 mm in diameter with a circular cross section. As a demonstration of MRI by the fabricated coil, we acquired MR images of an okra.

View full abstract

The communications tool is developed in the humanoid robot. The equipment development of the technology of miniaturizing technique is demanded because it becomes device arrangement at a position that is higher than center of gravity. Because the above-mentioned problem was solved by using the shape-memory alloy, it reports.

View full abstract

This paper reports the demonstration of a microrobot driven in the air powered by an insect muscule-powered bioactuator. Bioactuators are able to drive only in a culture medium. Therefore, we proposed a microrobot powered by the insect muscule packaged with culture medium in a capsule. First, we designed and fabricated the microrobot consisting of a frame and the capsule. Then, a dorsal vessel was assembled to the frame and put into the capsule. As a result, a leg deformation of 206 μm was obtained in the air. The results indicated the possibility to develop a legged locomotion microrobot powered by bioactuators.

View full abstract

In our study, we used the dorsal vessel (DV, heart-like tissue) of Drosophila melanogaster (Dmel, fruit fly) larva as a bioactuator. This paper reports on the closed loop visual feedback system to control the contraction rates of the DV in vivo. We investigated the response of the DVs that express channelrhodopsin(ChR2) and/or halorhodopsin(NpHR) to blue and red light. We found that DV expressing ChR2 increased contraction rates under blue light stimulation, while DV expressing NpHR stopped contracting under red light stimulation. In our feedback experiments, we managed to keep the contraction frequency of the DV in vivo above an arbitrary threshold value using the visual feedback system we developed.

View full abstract

In this study, we report on the blood pressure pulse wave device with prevention function for measurement position error by body movement. This device consists of a force sensor, a guide, and a acrylic plate for guide height adjustment. We measured blood pressure pulse wave by prototype device before and after ulnar flexion, and calculated the S/N ratio of blood pressure pulse wave. The S/N ratio decreased 30.1 % without a guide. When guide height is 0 mm, the S/N ratio decreased 2.9 %. Since the measurement position error by body movement was prevented by the guide, the decrease in the S/N ratio of blood pressure pulse wave was suppressed.

View full abstract

Most previous robots were designed to work in predefined environments and irregularities existing in the environment interfered with their operation. To tackle this problem, we focus on the locomotion of snakes that actively utilize terrain irregularities and move by effectively pushing their body against the scaffolds that they encounter. We have previously proposed a decentralized control scheme for a snake-like robot, which consists of only two local sensory feedback mechanisms. In this paper, we develop a snakelike robot and demonstrate through experiments that the robot with the proposed control scheme moves effectively by pushing its body against pegs.

View full abstract

Several animal species can move on various terrains, e.g., on the ground, in water, and in air, by adapting to the environment in real time. Numerous multi-terrestrial robots have been developed inspired by this ability of animals; however, their working area was still limited. To tackle this problem, we have previously developed an autonomous decentralized control scheme on the basis of a "scaffoldexploitation mechanism" inspired by flatworms, and have developed a two-dimensional sheet-like robot, SheetBot. Although the robot could move on various irregular terrains, it could not move in water. In this study, we resign the robot to move in water as well as on the ground. Experimental results show that the developed robot can move both on the ground and in water by obtaining reaction force from the environment effectively.

View full abstract

In this paper, we propose a discrete version of distributed building model inspired by wasp colonies. We define fundamental event rules in a cellular world, and introduce a action rule of distributed autonomous agent. Each robot decides its action rule following by the local configuration, without any communication between agents. Then we show that the agent build characteristic architecture in some rule.

View full abstract

In this paper, we propose a modular robot system adaptively transforming a mechanical structure with applying proximity sensors. A proximity sensor is consist of LED and a photo transistor. The modular robot is composed of nonidentical modules, easily replaced and connected. We make up autonomous decentralized control system by generating motivations of the modules to transform local structure with using proximity sensor output. In the first,we form a shapechangeable plane by connecting modules which have servomotors and proximity sensors. A module can estimate the distance between the module and objects by its proximity sensor. By equalizing the distance among neighboring modules, the modular robot system transforms along the shape of objects.

View full abstract

In this paper, we discuss some phenomena of distributed transportation inspired by foraging behavior of ants, in the light of space-discretization (or cellular automata) approach. We define fundamental event rules of autonomous agent model using two kinds of pheromones in the cellular world. We then introduce two types of ant agents: one is orienting ants, which follow pheromone trails, the other is random ants, which don't follow pheromone trails. First, we take a closer look at the number of pheromone trails formed by heterogeneous agents. Moreover, we propose a local feedback control method for the number of pheromone trails, and examine the effectiveness of the proposed method.

View full abstract

Most robots are vulnerable to the change in their own morphology such as failures. To tackle this problem, we focus on ophiuroids, which have simple body with pentaradial symmetry. Although they are composed of primitive decentralized nervous system, they can move by self-organizing their arm movements adequately even when the arms are arbitrarily cut off. As a first step, we concentrate on an ophiuroid with one arm and aim to clarify its locomotion mechanism. We propose a simple autonomous decentralized control scheme on the basis of behavioral experiments, and confirm its validity via simulations.

View full abstract

Snakes change their gait patterns in response to the environment. Elucidating the mechanism underlying this behavior lead to the development of robots that work in a wide range of environments. We have previously modeled the snake locomotion on the basis of autonomous decentralized control and demonstrated through simulations that the gait transition could be achieved only via a change in a small number of parameters. However, the previous model could not reproduce concertina and rectilinear locomotion, which are observed when snakes move in narrow space. In this study, we demonstrate through the simulation that concertina locomotion can be reproduced by an autonomous decentralized control scheme in which curvature derivative control is combined with a simple local reflexive mechanism.

View full abstract

It's always seemed like a big mystery: how nature, seemingly so effortlessly, manages to produce so much that seems to us so complex even if individual element or agent is very simple. This work was motivated by Swiss Robots, which collect scattered obstacles into some clusters without any global information nor intelligent concentrated controller. Then, we have focused on pattern generations, especially in static agent clustering by mobile agents. In this paper, we develop small mobile robots equipped with small motors to investigate pattern generations from experimental approach.

View full abstract