We propose a mapping and path planning method for 3D terrain. This method introduces Surface Mesh Map representation that has a graph structure. We use polygon mesh reconstruction method to generate the graph for path planning, which consists of nodes and arcs with costs of slope, roughness, height difference, and distance. The experiments show the proposed method is effective for path planning in 3D terrain.

The paper proposes an system for generating 2D drivable area map for an autonomous mobile robot in outdoor environment. We estimated the robot drivable area by Region growing segmentation based on normal information from 3D map generated with LIDAR. we detect large changes of state in the drivable area such as lawn from the road by using intensity information, and devised so that the robot can travel more safely. we generated 2D drivable map by merging these information. The experimental result show the 2D map correctly displays steps and obstacles for large-scale maps in urban areas and detect changes from road to lawn area. We performed an autonomous driving with an electric wheelchair using proposed 2D map. An electric wheelchair achieved autonomous travel in urban areas without collision with obstacles and invasion into lawn.

The method of searching the target within a known environment using a mobile robot has been proposed. In this paper, the target object is searched by detection using LRF and recognition using a camera. In order to improve the detection accuracy of the LRF, the environment is divided at a fixed area that the scanned resolution keeps the required resolution and searches are performed for each section. Also, in order to guarantee the minimum resolution of image recognition, a shooting point according to the performance of the camera is defined for the detected object. The mobile robot moves to a defined shooting point and shoots the object to specify the search target. The simulation result shows that the shooting point is correctly defined for the detected object.

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.

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.

We have been studying an air jet manipulation method to non-contactly carry an object over a long distance using multiple 3D air jet manipulation modules consisting of a single air jet nozzle and a pan-tilt actuator. Here we challenge a long distant transportation method thru object pitching and catching between a pair of air jet. In this report, we propose a control algorithm to determine each air jet angle and its flow rate for both pitching side and catching side. Several experiments on catching will be shown.

This paper presents the modeling and numerical simulation of a serial link robot fabricated by robot packaging method. First, we derive the equations of motion of a laminated robot made of a rigid serial link and a plastic film. We assume the plastic film as a closed-loop link structure with passive joints. The rigid serial link connected with a motor-actuated joint moves the closed-loop link structure (film) with passive joints. We numerically investigate the influence of the flexural rigidity of the plastic film on the motion of the rigid link.

In this paper, we develop a snake like fluid driven manipulator with curvature derivative control. With this control, the manipulator progresses forward when the lateral movement is constrained, otherwise every joints goes to be equal to a tip joint and the manipulator behaves like a elephant trunk. We propose a mechanism which reproduces a curvilinearly lateral undulation of snakes in 3 dimensions. In the manipulator, each joint is controlled by bellows actuation in response to slide valve in each link of snake body. We developed a prototype of the mechanism and experimented the performance of the prototype.

This paper presents an overview of pick and place operations performed by an inflatable robot. The inflatable robotic arm is composed of inflatable links and pneumatic bag actuators. This technology is expected to be applied to service robots working in human daily life, since the robotic arm has many advantages such as lightness, softness and safety. The pick and place tasks are realized by both pressure and visual feedback controls. The usefulness of the robot is demonstrated by several real world experiments.

This paper presents study of high safety redundant manipulator. Human friendly robot is increasing in recent years. But, they are often made of high stiffness actuator. For that reason, high stiffness actuator isn’t safe in case of collision with human. On the other hand, soft actuator is safe to contact because it is made of rubber and fiber. In this study, we use Pneumatic Rubber Artificial Muscles (PAM) as an actuator of the manipulators. Furthermore, we divide a manipulator into some units. At first, we manufacture the control of the unit and decide dimensions. Finally, we experimented on the drive of the unit that we made.

This paper describes a method for automated liquid manipulation by using a ferrofluid-based robotic sheet. The robotic sheet automatically moves a target liquid to a goal using the deformation characteristics of a ferrofluid and a visual feedback control with a camera. On the basis of previous findings regarding the flexibility of robotic sheets, we realized that our system could manually move a target liquid, such as water. In this paper, we elucidate the feasibility of the automatic manipulation and behavior of the target liquid during manipulation. Also, we quantify liquid mixing and elucidate the effect of the oscillation characteristics on this result.

We developed a compact size proximity sensor that detects the distance and the tilt angle of an object surface within 1ms, capable of the distance resolution approximately 20-times more precise than existing sensors. And, we realized a robust contact detection method using the sensor that does not depend on the reflectance, the tilt angle and shapes of an object surface. High-precision measurement of the distance enables contact detection independent of the contact force. In this paper, we describe the principle, parameters and the circuit of the sensor for precision measurement and contact detection method. We also demonstrated that the hand could catch the paper balloon fast without crushing it.

In recent years, robots are expected to do a variety of household chores. Among the tasks, we focus on daily assistance using bags because they are often used at everyday life. However, robots have difficulty in manipulating bags, which are deformable and difficult to see its inside during the manipulation. In this paper, we propose a solution for this problem by developing a gripper, on which proximity sensors are mounted on all sides. Use of this gripper enables robots to recognize the surroundings without contact even when they are unable to see it, and to manipulate objects in bags. We evaluate our method through an experiment in which a real robot picks an object from a flexible bag.

Robotic picking of folded objects such as books is required for picking various objects. To carry and place a folded object after it is unfolded, we propose a trial-and-error picking system using the Suction Pinching Hand, which can push the dangling part of the object until the object is folded. That system utilizes proximity sensors to detect the state of the object and predict whether folding will succeed. Also, proximity sensors, flex sensors and an air pressure sensor are used to deal with uncertainty of the image recognition, the hardware and suction grasp. We confirmed with an experiment using a real robot that our proposed system can carry a folded object stably and place it in narrow space.

Human hand can both exert large gripping force and grasp various objects dexterously. However, it is difficult for most robot hands to manage both. To develop such hands, we focused on the thumb CM joint with wide range of motion and the MP joints of four fingers with the DOF of abduction and adduction. We developed the strong and dexterous hand by applying the wide range joint and joint mechanism with variable stiffness to the hand using machined springs, which can support large load because of their flexibility and robustness. We realized the grasping of various objects, supporting large load and several motions with the developed hand with an arm.

With the development of robot technology, the expectation for humanoid robots working in the same environment as human beings is increasing. In order for humanoid robots to realize various tasks as human beings, robot hands corresponding to human hands Function is indispensable. In this paper, we design and produce a hand robot hand that can reproduce various human grip methods. In addition, we constructed a position control that allows the hand motion measured by optical motion capture to follow up, and constructed a force control that adjusts to an appropriate grip force. And we constructed a hybrid control system combining them. Furthermore, we verify the effectiveness of the control system using the robot hand.

Articulated mechanical hand is a robot hand for industrial robot which has enveloping grasp capability without sensory feedback and can grasp various shaped and different sized objects mechanically. In this research, a geometrical configuration between hand and object for a suitable enveloping grasp and the equilibrium between joint torque and weight of object keeping the grasp stable are considered and the enveloping grasp stability of the hand is studied.

The purpose of this research is the sharing of mapping data on multiple units for environment mapping by the SLAM method and improvement of mapping accuracy. I assume robots are to work in the same room as the person. I thought that there were a lot of obstacles in the same room as persons, and it was an environment with a dynamic obstacles.

The robot needs to become small to work in narrow space. In addition, the robot needs to decide the information from the external sensor attached to the robot itself. working one room by one small robot when creating a map is inefficient. So, I thought about mapping and sharing with multiple robots. Because I thought that making of maps is possible efficiently by sharing data. I did experiments on environment mapping using a small sensor of a small mobile robot and examined its accuracy. In addition, as a result, we obtained accuracy of robot's self-position estimation and accuracy of a small distance measuring sensor used for environmental map construction experiment. Then we were able to know about these errors and remedial measures.

In this paper, we discuss a system in which the base with a flexible beams moves including contacts with the ground. We propose a feedback control method for the flexible beam to support the locomotion of the base part. Due to the complexity of the contact model with the ground and the high degree of freedom of the flexible beam model, Distributed Model Predictive Control is adopted for the proposed method. For our purpose, we consider that the state in which the force/moment applied to the base of the flexible beam faces the direction in which the robot should move is good. The control input of the flexible beam that maximizes the degree of coincidence between the direction of the force/moment applied to the base of it and the direction to the target state of the robot's body is obtained.

Despite the numerous advantages associated with the flexible manipulators, link vibrations stand in the way to reaping these benefits. This leads to time wastage waiting for vibrations to decay to safe operating levels and the possibility of mechanical failure due to vibration fatigue. This paper presents direct strain feedback vibration control by tuning the feedback gains using artificial neural networks on a 3D flexible manipulator. Online backpropagation was developed in MatLab Simulink and implemented in dSPACE environment for practical experiments. Results show significant reduction in the link vibration relative to the performance of fixed feedback gain.

This study proposes a gripper which has compliant finger mechanism for universal manipulation systems. Conventional automation systems usually utilize customized peripheral devices for a target object to keep the relative position between a robot and an object. These devices cannot be used for other objects that have different sizes and shapes, which makes such systems not universal. To achieve a universal automation system, it is necessary to use universal devices instead of customized devices; however, universal devices do not have the complete positioning function, which may cause unexpected collisions between the robot and surroundings. To deal with this problem, we focused on the mechanical compliance of end-effecters. We developed a two-finger gripper with a passively compliant mechanism which combines a parallel four-bar linkage and a slider-crank linkage. Finally we showed the effectiveness of the compliant function of the gripper by picking-object experiments.

Bag-type grippers capable of grasping objects of various shapes have been developed so far. However, the conventional bag-type gripper mechanism has a problem that a large sliding resistance is generated due to contact with the surrounding environment during insertion into narrow space. This large sliding resistance causes collapse of the surrounding environment and breakage of the gripper itself. The authors propose the dual layer type torus gripper mechanism with extremely low sliding resistance in this paper. We developed the prototype model of the dual layer type torus gripper mechanism and verified its basic performance by some experiments.

By imitating the hands of a tree frog, a method to improve the performance of adsorption gripping is proposed. By using a robot hand equipped with multiple small suction cups like the frog, an object having complicated curved surface could be gripped. Grasping experiments were carried out, where the performances of a single large suction cup and multiple small suction cups were compared with. It was proven that multiple small suction cups are effective, even if there exit suction cups not contacting with the object surface. It is because the vacuum leak amount is small thanks to the smallness of each suction cup, which is also effective for eliminating comparative large and expensive check valves from the robotic hand.

The surface of robotic hand requires high friction for stable grasping. A soft material, e.g. silicone, was popularly used for the surface owing to its high friction. However, a slippery surface is also required in some situations such as inserting fingers into a narrow space. In such situations, a grippy surface could prevent a smooth operation. Additionally, the operation is not always done in dry condition. Therefore, the aim of this study is to develop a soft robotic fingertip with high friction surface under both dry and wet conditions while to develop a friction reduction system based on lubricating effect. The efficacy of the proposed robotic fingertip composed by the high friction surface and friction reduction was examined in actual tasks.

Cooperative manipulation with multiple position-controlled manipulators is a very attractive technique to enhance the applicable fields of industrial manipulators. However, excessive internal forces occur among manipulators caused by mutual positioning errors. So we proposed a way of installing passive joints between each manipulator tip and the handled object to avoid the excessive inner forces in our previous works. The system works very well, but the application of the system is limited to the cooperative transportation. To enhance the fields of our cooperative system with passive joints, it is proposed to install coil springs into each passive joints to generate inner forces among manipulator tips. By installing springs into passive joints, it can be applied to tasks requiring forces between manipulator tips and the handled object. In this paper, the relationship between displacements of springs and the generated inner forces is derived and the statics of the cooperative system is analyzed by fundamental experiments with a prototype mechanism.

Tele-operation system is required to provide reaction force in a remote robot to its operator. The authors have developed a master-slave system which consists of a multi-fingered haptic interface and a humanoid hand robot. Both the haptic interface and the robot hand have 6-DOF robot arm and 15-DOF robot hand. This research is aimed for improvement of operability of the master-slave system. The master side control system is designed by using adaptation control gain. Experiments show the effectiveness of the proposed method.

We developed a one-dimensional regulated liquid level stereolithography method, which is a new 3D printing technology that is compatible with definition and modeling speed.

In this presentation, we describe technical explanation and features of this method and examples of stereolithography that can be applied to medical treatment.

Transsphenoidal surgery requires the surgeon to insert surgical instruments through the nasal cavity to the brain tissue and to perform fine manipulations. Introducing surgical robots into transsphenoidal surgery potentially provides various benefits such as reducing hand tremor and increasing precision, while the lack of haptic feedback remains as one of the main technical challenges. To overcome this issue, we introduce a miniaturized (3×2×0.5mm) force sensor that can be introduced into surgical robots. In this paper, we introduce the overview of the structure and its preliminary evaluations.

Blood vessel models for medical training have been developed to represent realistic geometries and mechanical properties of the realistic blood vessels. However, few models were developed for assembling the other organ models. The purpose of this study is to develop a new blood vessel model which can be mounted on the Bionic Humanoid. The geometries and the mechanical properties of realistic cerebral arteries were represented using poly (vinyl alcohol) hydrogel (PVA-H) and silicone. We observed the PVA-H model can be ruptured depending on the force on the catheters and the guidewires during operations. This result cannot be observed using the previous blood vessel models produced from silicone or the other materials, and this observation indicates the PVA-H model can represent the realistic situations in operations.

Suturing task is one of the most technique-demanding and time consuming tasks in the endoscopic surgery. We have developed a novel needle driver by that this task can be easily preformed with a simple action. Before this device, only 10mm devices were available. By proposing a simple wire driven mechanism, the novel device have a shaft which is 5mm in diameter. With a customized straight needle, the device can puncture the target tissue with the needle continuously in the patient body. This paper presents the mechanical details of a prototype of this device, and feasibility showing how it works.

In recent years, researches are being carried out on the creating of surgical simulators for brain subjects, aimed at surgical training of young physicians. Also, based on the creating of the brain surgery simulator, not only the creating of tissue models such as dura mater, brain, and tumor, but also their evaluation becomes important. Living tissue has both elasticity and viscosity, and by evaluating both, it will evaluate the mechanical properties.

In surgical skill assessment for pediatric endoscopic surgery, expert surgeons review surgical videos to evaluate, which is cumbersome and time-consuming. The present authors propose to automatically assess pediatric surgical skills by extracting the surgical needle in videos. This study presents an image processing method to extract a surgical needle in surgical simulation videos where synthetic tubes simulating the esophagus is endoscopically anatomized in a pediatric chest model. The proposed method was evaluated using a surgical simulation video, and the accuracy for detecting the needle insertion point was 19 pixels, which was equivalent to 0.56 mm.

An interesting phenomenon in which a vortex flow occurs in a circle chamber is observed when we impart periodical pressure vibration in the main channel where the circle chamber is connected at the corner with sharp bending angle. This phenomenon is called Vibration based Virtual Vortex Gear (V³G). A vortex flow can be induced under V³G without continuous fluid flow at the entrance of the chamber. By the characteristic, it is possible to mix micro objects in the chamber with V³G mode. Furthermore, the characteristics of V³G is enhanced by an additional air chamber. We developed an experiment system and evaluated the distribution of beads density by using luminance level for evaluating the effect of V³G.

Smart contact lenses are next-generation products that allow a lot more than just perfect vision. In 2014, Google showed a proof-of-concept electrical contact lens which assists people with diabetes by monitoring the glucose level in their tears. Other research efforts include prof. Brian Otis at the University of Washington, who made a readout IC chip for wireless communication from a contact lens. Separately, prof. Shelley D. Minteer at the University of Utah developed glucose sensors and biofuel cell systems on contact lenses. However, all such attempts apply dry lithography on dry contact lenses, even though most people prefer to wear moist soft contact lenses. This is because most electrical circuit printing techniques are difficult to use on soft, moist surfaces. Here we demonstrate a direct bonding of wireless circuits on moist soft contact lens with our original electrochemical (EC) bonding technique. The EC bonding allows a washable and wireless powered contact lens at 10mm distance between the electrical eyeglasses and the designed contact lens. The power from the eyeglass sources was transferred at around 20 % to the electrical contact lens when the frequency is coupled at 13.56MHz.

As biotechnology develops, sophisticated manipulation techniques such as moving arbitrary cells to arbitrary positions are required. Conventionally, there is a method of moving each cell by bringing it into contact by using a thin needle. However, this method has various problems such as damage to cells, efficiency, technical aspects. In this research, we aimed to manipulate arbitrary cells noncontactly by vibration, examined the mechanism, verified by prototype and experiment.

In microscopic field, measuring a depth distance and a direction of a focal plane on an optical axis is very difficult. In our previous study, the depth estimation method are proposed based on difference between blur widths of objects in a single shot. However, the direction of a focal plane is unknown by the method. In this study, we propose an effective approach to detect the focal plane in microscopic field. We focus on different corrections of chromatic aberration to detect the plane. The corrections is roughly classified into four types. These types show different focus shifts for each light wavelength. Therefore, color appearance on the blur width of a target differ from each types. Extracting the color appearance by image processing, we detect the focal plane. From experimental results, we showed the effectiveness of the proposal approach considering focus shifts of chromatic aberration.

Molecular delivery into the cells is an attractive technique to functionalize or analyze the cells in medical biology. Recently, we have developed Au nanostraws in track-etched polycarbonate membrane with gold electroless plating, and subsequent chemical and dry etching. Here, we demonstrate the molecular delivery into adhesive cells using Au nanostraw stamping. After the stamping, we measure both the reliability (live/dead cells ratio) and delivery rate (stained cell number / total number).

In recent years, studies of microrobots have been actively cunduted. However, little microrobots achieved complicated movement like living things. We aim to realize a microrobot capable of complicated movement mimicking living microorganisms. In this study, we propose a microgel robot that can perform peristaltic motion. The proposed robot mimicking structures of microorganisms, that has hybrid structure of soft and hard materials. As a hard material, SU-8 which is hard photoresist is used. As a soft material, bioresist which is a temperature-responsive gel is used, and a light absorber is mixed with it. It enables to actuate this soft material by light irradiation. By using dextran, a polysaccharide, as a sacrificial layer material, we succeeded in patterning microgel robot and peeling from Si substrate.

A surgical robot for vitreoretinal surgery has been developed, specifically, for the peeling of ILM (Inner Limiting Membrane) and cannulation to a blood vessel in the eye ground. We have been studying the automation of robotic vitreoretinal surgery to enhance the accuracy and safety. We proposed a method to autonomously detect the height of the robotic tool tip using its shadow projected on the eye ground. In the proposed method, the positions of the shadow is measured by changing the position of the light source, and the height of the robotic tool is estimated to be used for its autonomous positioning in the depth direction. In this paper, we present a method to enhance the accuracy of our proposed method by increasing acquisition points of the tool-tip shadow. For ten trials for the estimation of a 5 mm height, the average error for the height estimation was 0.272 mm.

In this paper, we develop a brain model to be incorporated into a surgical simulator for training of brain surgery operation. In order to mimic the physical properties of the actual human brain, a brain model made of polyurethane gel imitating the shape of the brain was prepared. In order to select materials close to the brain, dynamic viscoelasticity test was conducted to compare the viscoelasticity distribution of porcine brain, urethane gel, urethane gel and salt compound. As a result, it showed that the viscoelasticity distribution of urethane gel with salt concentration of 10% is closest to porcine brain.

As the medical robotic systems evolve, becoming more sophisticated in their working, it is possible to assist minimally-invasive surgical procedures that otherwise could be impossible or risky using traditional approaches. Despite the evident benefits of these robotic systems, there are many open problems that must be solved to guarantee the safety before the commercialization and introduction of these systems into the surgical room. In the other hand, the use of virtual reality (VR) simulators may play an active role during the development of the robotic systems. In this work we explore the use of VR simulation applied to robot-assisted transsphenoidal tumor resection, to incorporating collision avoidance and posture correction to an articulated tool, the results suggest a decrease in the number of collisions by adding constraints inside the VR simulation, which are solved as part of the interaction flow.

A surgical simulator for operations is used for improving doctor-s’ surgical techniques. The surgical operation simulator has the mainly two types: an actual model and a virtual model. The surgical operation simulators using a model can use the operation apparatus in actual surgical operations, so doctors can practice operations in almost the same environment actual operations. This study makes dura mater models for the surgical simulation of minimally invasive brain surgery. This paper shows material characteristics, fabrication method of artificial dura mater model, evaluation of fabricated dura mater model, and simulated surgical evaluation of fabricated dura mater model.

In our previous work, we have proposed a hands-on user interface designed to provide a surgeon with intuitive and safe operation of articulated forceps based on a passive mechanism. In this paper, we propose a spring balance-based gravity compensation mechanism for our proposed user interface to improve its operability. We present experimental results to demonstrate the balancing performance and usability in a tracing task.

This paper describes a brain retraction pressure sensor using a fiber-optic pressure sensor for neurosurgical application. The requirement for this pressure sensor is: it can measure a pressure of 60 mmHg; it can be attached to arbitrary place on a brain spatula; and it can prevent electric leakage, can be unaffected by electromagnetic noise. It can also do temperature compensation. The pressure sensor was made of a fiber-optic pressure sensor and thermocouple placed in a concave of a metal part filled with silicone rubber and it was mounted on a brain spatula with steri tapes. Pressure measurement and temperature characteristics were evaluated. In the evaluation result, the sensor output was linear with the applied pressure, and it could do temperature compensation. Retraction test to the brain model was performed using the brain spatula equipped with the pressure sensor. As a result, doctor could confirm the exclusion pressure value.

Vascular network in cell-dense structure is indispensable to avoid necrosis of cells densely located. In this paper, we propose a fabrication method of 3D vascular network in cell-dense tissue using magnetic fibers. Especially, a hepatic lobule-like vascular network is constructed by manipulating magnetic fibers. The hepatic lobule shows a hexagonal shape and multi-layered structure. For this, steel rods connected with the magnetic tweezer are used as portal and central veins, in addition, fibers play a role of sinusoid that transfers nutrients from portal to central vein. To attract and assemble magnetic fibers, enough magnetic fields are required and then we design a magnetic tweezer with seven poles based on the simulation data. Analysis of magnetic fields is attained by 3D electromagnetic simulation software, EMWorks. Experiments are conducted by using the proposed system, and therefore, we construct three-layered hepatic lobule like structure in fibrin gel.

Conventional actuators such as geared servos suffer drawbacks in utilizing them in robots which work closely with humans. Dielectric elastomer actuators (DEAs) overcome these challenges since they can realize natural motion patterns which are inspired by those of human muscle. To date, fabrication of such bio-mimetic soft actuators is hindered by limited electrical conductivity of the electrode materials. We aimed at developing highly conductive multi-wall carbon nanotube/polydimethylsiloxane (MWCNT/PDMS) composites for use in DEA electrodes and to develop a simple single layer actuator fabricated with these improved composites. To this end, we proposed a method to increase the dispersion of MWCNTs in PDMS and to increase the electric conductivity of the composite. Our results showed that simple crushing of the composites in addition to the use of naphthalene/toluene improved the dispersion of MWCNTs in PDMS and enhanced the DC electric conductivity of the composite. We also proposed a simple conceptual design and developed a single layer DEA based on these improved composites.

Our photochemical cell membrane perforation method is new approach for mass cell processing, eg. cell therapy or regenerative medicine. The method requires homogeneous control of cell attaching pressure of nano brush shaped cell membrane perforator which contains photosensitizer, but it was hard because almost the cell size are micrometer scale and its structure is weak. We made the fluid-packed perforator to homogenize the applying pressure, based on Pascal's principle. This perforator has flexible surface which can adopt variations of cell layer on culture dishes. In this research, we developed the extended system for tissues and will report a trial effort on mouse tissues.

Tumor specific binding peptides have the potential to alleviate the side effects of anti-cancer drugs. To reduce the laborious effort for peptide screening, some microfluidic devices were proposed using adherent cancer cells or protein coated beads as the targets of peptides. We have developed a microfluidic peptide screening device with a group of cancer cell spheroids as the targets of peptides, which are closer to in vivo tumors. Channels are designed based on the fluid behavior to trap minute micro scale spheroids in the chambers of the device. The experiment results demonstrate that the fluid behaves as designed and the device can successfully form a group of spheroids.

In this paper, we propose a stacked rotary actuator, which is a high power actuator driven by molecular motors. By stacking rotation units driven by molecular motors, the number of molecular motors working cooperatively increases and we expect to realize a high power actuator. In order to fabricate rotation units, it is necessary to control the orientation of the molecular motor. In this study, orientation control was performed using SU-8 microstructure. For stacking rotation units, we designed the microstructure which can fabricate many rotation units at once. The results confirm that the molecular motor were circularly oriented along the wall surface of the microstructure.

We created bio-actuated piston micro pump using earthworm as a bio-actuated micro piston. We first designed cylinder channel that earthworm reciprocate to move fluids. It is prepared with micro fluidic chip. Then we demonstrated fluid actuation by movement of earthworm in cylinder channel. We estimated the flow rate from tracking the micro spherical beads. The flow rate was 0.23 µL min^-1. It was equivalent to current bio-actuated micro pumps. Finally, we show the driving of the valve by earthworm in micro fluidic chip with planar check valve. The check valve was moved 31.4 µm. It was found that the bio-actuated micro piston pump had high space efficiency and high pressure potential.