Recently, multiagent systems has attracted much attention. RoboCup Soccer Simulation is treated as a test bed of these. In real soccer, the analysis of games in terms of team behavior has been studied, however, it has not been done in RoboCup Soccer Simulation. In this study, we aim to facilitate the analysis of the team behavior and to clarify the involvement of the team possession type in result of games in RoboCup Soccer Simulation. In particular, we construct a method of detecting counter attack, one of team possession type and we analyze the log files of games. After considering the feature quantities to detect, we detect counter attack using support vector machine based on four feature quantities. As a result, this detector can detect in 77%. Using this detector will reduce the burden of check the log data in visual.

View full abstract

We have studied a giant swing robot whose giant swing motion can be obtained simply by applying Q learning method as one of reinforcement learning methods. In this study, we focused on learning generalization that is a problem whether a robot can choose one rotational direction. In our previous studies, it was successful that the real robot can't acquire backward rotation but forward rotation. This paper clarifies whether learning generalization can apply both forward and backward rotation on the simulation base. Learning simulation experiments take place in the various conditions like rewards, actions after rotations and learning system. As a results, it becomes obvious that trend of rotational direction can be controlled to some extent.

View full abstract

We clarify design principle, i.e., emergent mechanism of distance measurement function, by using evolutionary robotic system which designs morphology and intelligence autonomously. So far, we have shown the distance measurement function by contacting own body structure such as a tactile sensor in case of short range and by stereovision using sensor arrangement in case of middle range. However, the emergent function in case of long range is still unexplained. In nature, a predator measures long distance to a prey accurately by using the memory. Thus, we gave the robot with memory a task keeping distance to a target and investigated the emergent function in case of long range. As a result, we clarified following two design principle. In case of long range, the distance measurement function by using parallax between current and previous sensor conditions emerged. Furthermore, the long range task simplified the morphology and developed the intelligence using memory.

View full abstract

In this paper, we clarify relationship between spatio-temporal resolution of a visual sensor and body size of a robot using an evolutionary robotic system that designs its own morphology and intelligence simultaneously. In nature, it tends to be that the smaller the creature size, the lower the spatial resolution of the visual sensor and the higher the temporal resolution. To investigate this mechanism, we gave robots with different sampling frequency of the visual sensor, number of the pixels and body size a task of maintaining a certain distance to a target and performed evolutionary simulation using genetic programming (GP). The result showed two design principles: the higher the spatio resolution, the better the stereo vision strategy with high frequency and the pursuit strategy with low frequency. Moreover, the higher the spatio resolution, the more downsized the body size with stereovision strategy.

View full abstract

In this study, we aim to evolving autonomous virtual creatures which have various shapes in a complex environment. We implement a basic physics law and fluid influences with a virtual environment and evolve virtual creature in various environments. We adopt central pattern generator as controller of the virtual creature. Each model obtains an effective moving behavior in physical environment.

View full abstract

A multi-robot system is composed of multiple robots which are relatively simple. In this system, reinforcement learning is one of promising approaches for controlling each robot. However, its performance depends a great deal on the segmentation of state and action spaces. To deal with this problem, we have been developing a new technique, named BRL. This paper introduce a meta-learning mechanism to standard BRL in order to improve its learning ability. We investigate the performance of extended BRL through physical experiments. A task is that mobile robots orbit an object in an environment.

View full abstract

An environment plays an important role in behaviors acquisition for artificial creatures. Thus, the environment must obey the physical laws. In this paper, it is examined how the behavior differences appear when the artificial creature with wings, autonomously behaves in the air and water environment. We propose a simulation method for the artificial creature using evolutionary computation in two fluid environments with low computing cost. As a result of simulation, we verify that it is possible for the creature to acquire adaptive behaviors in different environments. After evolution, the creature behaves autonomously by leveraging effectively fluid forces in each virtual environment.

View full abstract

Computation methods of parallel problem solving using graphic processing units (GPUs) have attracted much research interests in recent years. Genetic algorithms (GAs) can be implemented to GPUs in terms of the parallel processing of individuals in a population. This paper describes yet another implementation method of GAs to the CUDA environment where CUDA is a general-purpose computation environment for GPUs provided by NVIDIA. The major characteristic point of this study is that a steady-state GA is implemented on GPU utilizing on-chip memory in order to solve difficult optimization problems. The proposed implementation is evaluated through four test functions, then we found that the proposed implementation method yields 5.2-14.8 times faster results than those of a CPU implementation.

View full abstract

Probabilistic Traveling Salesman Problem (PTSP) is used to study routing problem under uncertainty. The PTSP is similar to the TSP with the difference that each node has a probability of requiring a visit. Optimization approach for the PTSP consists in finding optimal solution that visits all the nodes and that minimizes the expected cost of the solutions. But, this problem becomes very difficult when the numbers of the node increase. In this paper, we apply LCO and GA into the PTSP to solve this problem and their performance is illustrated based on the experimental results.

View full abstract

This paper proposes a new Q-learning which can construct the mapping function between the continuous state and action space. The proposed method can calculate the expected values of each action by using neural networks and CMAC which is one kind of tiling technics. The proposed method is evaluated through a navigation problem of an autonomous mobile robot.

View full abstract

A cat can land safely on the ground from any posture in free fall with no initial angular momentum. To right itself, the cat does complex motion to overcome the contradiction between the rotation of its body and the conservation law of angular momentum. This study explores how to control the cat posture in a space based on physics modeling and evolutionary neural networks.

View full abstract

Many researchers are analyzing human networks. It is thought that such a movement will become popular more and more in a future, because human network analysis is useful for social network service(SNS) and management strategy of corporations, etc. We identify a preferable structure of human network to find methodology of rebuild an ideal social organizations with the property value of the network. Using the property value, we rebuild a human network with a previous human network by simulation.

View full abstract

An elastic robot is proposed which has a body with multiple degrees of freedom. This study focuses to allow the elastic robot to behave autonomously. A physical simulation is implemented to achieve given tasks for the virtual elastic robot. We assume a complicated task to acquire locomotion in which an obstacle is placed. We conduct a parameter optimization of control system by Evolutionary Computation. Moreover, in order to design an intelligent behavior, this paper proposes "Behavior Composed" which consists of several simple behaviors. Experimental results show that the robot is capable of acquiring locomotion in specific circumstances.

View full abstract

The Loosely Symmetric model (LS model) is a model of human illogical decision. A previous study revealed that agents implemented LS model achieved high performance for the N-arm bandit problem. In this study, we evaluate the performance of LS model applied in Q-learning, using the Acrobot example.

View full abstract

Learning strategy (LS) fusion is our previous work in reinforcement learning (RL) framework. LS fusion fuses multiple LSs, such as transfer learning, for a single task of a robot. This paper introduces two LSs into LS fusion: an LS to learn a dynamics and a reward models, and an LS using a model-based RL method. Especially, we propose to use the MixFS dynamics model which is also our previous work. MixFS decomposes the dynamics model into the task specific elements and the task invariant elements. Thus, we can initialize the dynamics model of a task by transferring the one of the other task. In simulation experiments, we apply LS fusion with the new LSs to maze tasks of a small humanoid robot, where the primitive motions, crawling and turning, are also pre-learned by LS fusion. The results demonstrate that the new LSs improve the learning speed by using MixFS.

View full abstract

Reinforcement Learning (RL) is one of promising approaches for controlling an autonomous robot. However, its performance is quite sensitive to the segmentation of state and action spaces. To deal with this problem, we have proposed one RL technique, named BRL. BRL has the feature of segmenting state and action spaces autonomously. In this paper, we investigate the performance of BRL through cooperative Box-Pushing task using four mobile robots. Moreover, for treating that tasks, we adopt Shaping which is the notion extensively applied in various fields. Shaping make robot learn efficiently while raising difficulty step by step. It is considered effective to deal with task, because robots can search with previous learning. In addition, the cooperative behavior which is acquired by using Shaping is analyzed.

View full abstract

Recently, robots are expected to work around people like care robots and robot pet. Robots to work around differences in industrial robots and humans, whereas the former was assumed to operate in places that the robot activity, the latter is primarily a human activity is not expected activities of the robot It is a place to work. Robots used around human need to correspond to a sudden change in behavior and to respond as human beings and the environment. In addition, make robots work around the human need to build a computational model to simulate the motion of the robot for a wide variety. In this study, constitute a computational model that can respond to changes in kinematics for multi-agent concept, and develop computational tools of the movement.

View full abstract

We, in this study, try to acquire cooperative behaviors of multiple robots on the basis of evolutionary robotics approach. As the cooperative task, the 3-Dimensional balancing problem is given. This problem is an example of corresponding tasks to find the appropriate movements for multiple mobile robots. Each robot agent on a flat board keeps being balanced state by using evolutionary robotics approach. The behaviors of mobile robots are calculated by corresponding neuro-controller which consists of artificial neural networks. Evolutionary computation method is applied to evolve weight values of neuro-controller. Numerical simulations are carried out by a simulator constructed based on 3-D physical engine ODE(Open Dynamic Engine). As a result of simulations, usefulness of our approach is clarified.

View full abstract

We propose a method to genarate sight feature space for motion acquisition of robots The sight feature space is generated by embedding a vector composed of the points of features of the image to a low-dimensional space using LLE(Locally Linear Embedding). LLE is a manifold learning method for distribution of high-dimensional data to a low-dimensional space using its local linear structure. Image feature is extracted from the image including the arm of robot and the object using a SIFT(Scale Invariant Feature Transform).

View full abstract

A method which co-evolves genotypes of the behavior and morphology is effective in creating the artificial creature. Specially, the diversity of the morphology is important for use of evolutionary computing, because the morphology needs to change itself due to a given task. In the paper, we aim to create the best artificial creature by using evolutionary computing in 3D physical surroundings. Specially, we focus on fluctuation of phenotype.

View full abstract

This paper presents a learning approach to acquire holding-up manipulation by a humanoid robot. Information of force sensors attached at the surface of the arms are mapped to a lower-dimensional space and utilized to find a better configuration for the manipulation task. The proposed motion generation framework was verified by simulations with a humanoid robot. It was shown that the robot could judge failure of its manipulation using force sensor information, and succeed to modify its configuration so as to realize holding-up manipulation.

View full abstract

A multi-robot systems (MRS) is composed of many robots. So far MRS has been studied in various fields. For a robust MRS, we focus on Reinforcement Learning to control a MRS. Reinforcement Learning works well in a static environment. Since, it is difficult for Reinforcement Learning to adopt a dynamic environment. In addition to that, in general, a designer needs to appropriately decide how the discretization level of the state and action spaces is represented. However, there is no general design guideline. To overcome these problems, we have been developing BRL as a type of Reinforcement Learning techniques. In this paper, a technique is proposed for effectively using messy knowledge acquired using BRL for improving the robustness of MRS. The technique reconstructs the state space by using Support Vector Machine based on the input-output data acquired by BRL. To investigate our proposed technique, We conduct computer simulations of a cooperative carrying task with three autonomous mobile robots.

View full abstract

In this paper, we study a method to translate the style of an input motion which is obtained by a motion capture system. For this aim, we model the difference of motions coming from the actor's attribute, such as gender and age. Especially, we focus on the difference coming from the gender, and model the difference of poses (joint angles) between males and females by using a Gaussian process regression. In training the model, we utilize Multifactor Gaussian Process Model to align the training set of motions in cyclic phase. In the experiments, we train the filter with walking motions of males and females, then we apply the filter to walking, running, and jumping motions of males. The results demonstrate that the filter successfully translates the motion into a womanish one.

View full abstract

In this study, we choose reinforcement learning as a method of acquiring a robot's behavior autonomously and aim at walking behavior acquisition with a gecko-type, four-legged robot with a waist joint. The setting of the reward function has great influence on the learning result because the aim of the agents in reinforcement learning is to maximize the total acquisition reward. We performed an experiment in which a straight line was set to the target orbit and an error between the generated orbit and the target orbit was added to the reward function, and examined how design of the reward function influenced the generated orbit. As a result of the experiment, our robot acquired efficient walking with a waist joint, so that error between the generated orbit and the target orbit was minimized while moving forward.

View full abstract

Many real creatures have muscles to control their flexible movement. This study proposes a virtual muscle model and its control method for artificial swimming creature. The virtual muscles are composed of elastic springs for the uses of actuators. ANN controls these actuators. The virtual muscle's parameters and ANN's parameters are optimized for the creature to achieve proper swimming by EC and PSO. Using simulation results, the swimming of the artificial creature can be compared to the real swimming creature.

View full abstract

This study aims at the development of mobile rovers for conservation or investigation of tidal flat. Especially, this paper focuses on wheeled rovers which have the simple structure. Wheeled rovers have high possibility of tire spin because of tidal flat covered with mud. In order to avoid tire spin, three flexible wheels are adopted for the experiments on the difference of flexure. From the experimental results, the most flexural wheel shows the low slip ratio on the imitating tidal flat using muddy soil.

View full abstract

Effects of body extension and its relationship to learning in robots is not clear even though lots of research have been done. In this paper, to reveal these characteristics, we make simulation robots and attach some rigid bodies to them in order to extend their bodies. Through the interaction in game, extensions increase the average winning percentage of some rule sets. But it does not affect the winning percentage of a specific rule set. Also, each extension has its own collision property and its average collision time is not proportional to the surface area of extensions. We use simulation environment which cause physical effects to robots by using physics engine.

View full abstract

We have modeled load carrying robots by applying the physical engine (PhysX) to a virtual environment. This study aims at acquiring an autonomous and cooperative behavior for load carrying robots which moves from a depot to a destination place through the virtual force propagation. The load carrying robots are controlled by ANN (Artificial Neural Network), which is evolved by learning proper actions by use of GA (Genetic Algorithm). Simulation results show that the load carrying robots acquire autonomous and cooperative behaviors under the virtual physical environment.

View full abstract

In automatic composing system, the system has to generate a music that matches to the user's preferences and favors. To solve the problems it is necessary to get parameters for numerizing and estimating user's preferences and favors. As related works, some interactive methods by which parameters and got ea from user's response were proposed. Some system also researched that automatically extract characteristic values from sample musics. In this paper, we propose an automatic composing system that generate melody matching to user's favors. This system consists complex network as a melody generation engine, and artificial neural network as a melody evaluation engine. By this system, we try to generate and evolve a prefer melody from sample musics.

View full abstract

We propose a simple physics model of a butterfly and its flight control by the real-coded genetic algorithm (RCGA) and the artificial neural network (ANN). A physics model consists of two kinetic equations which are led by simplification of the fluid force. A butterfly's flight is controlled by an ANN. The RCGA optimizes weights of the ANN for obtaining the suitable flight. After evolution, the generality of the optimized ANN is investigated by changing an initial height at which a butterfly starts in flight. Then it is shown that the flying creature enough evolved can not keep at the aimed height when the initial height is set far from the initial height used in evolution. Investigation by changing generality of the ANN in evolution shows that too much optimization may reduce the generality of the ANN.

View full abstract

We have proposed technique, named Bayesian-discrimination-function-based Reinforcement Learning (BRL). BRL can segmentalize continuous state and action spaces. Because of this, we succeeded acquisition of cooperative behavior in Multi-Robot System under a real environment with noise and uncertainty. In that experiment, robots didn't understand who their partners are. So in this paper, we aim to acquisition of cooperative behavior between robots and a human by BRL. A task is that robots keep a pallet upright and lift up it with a human. We examine the robustness of BRL through the change of an environment by shift to another human.

View full abstract

Many studies have been done on autonomous locomotion robots based on physics modeling. However, most of them concern on walking and swimming locomotion and there are few studies on a series of walking locomotion. In this study, we focus on a dog walking locomotion and on dog's standing up behavior, which is a starting behavior of walking. We applied GA and ANN to acquire the standing up behavior based on physics modeling and verified that a suitable behavior can be acquired through numerical experiments.

View full abstract

This paper attempts to acquire locomotion patterns of locomotion robots by using Q-Learning method. Generally, primitive motion forms are achieved using basic motion rhythm by designers. However, in this study, locomotion robots voluntarily acquire locomotion patterns only by using information of interaction with environment. An advantage of this method is able to ensure that these motions are best for these robot mechanisms and environment they exist. In this paper, we apply reinforcement learning to some six-legged robots which have different shapes respectively in simulation. Then we compare each locomotion patterns and find an actual insect locomotion pattern by clarifying mechanical factors.

View full abstract

In this paper, we newly propose an ultrahigh-speed sensing and control platform to manipulate a single cell with high-speed. The developed system is composed of an XYθ high-speed stage to drive a 3DOF-MMT (Magnetically Driven Microtool), an on-line high-speed vision capable of 2000 frames/s image capturing with 512 x 512 pixels, and a high-response syringe pump by using a high-performance AC motor. The specific design and signal flow of the system are also shown. The frequency response characteristic of the MMT is confirmed using the developed platform with a fabricated microfluidic biochip. Finally, high-speed particle positioning in a microchip is achieved by the combination of the developed pump and the on-line vision.

View full abstract

This paper presents the comprehensive analysis of fluid force, the optimum design and its fabrication to reduce the fluid force on the magnetically driven microtool (MMT) by attaching riblet shape on the MMT. Ni and SU-8 composite fabrication is employed to form three-dimensional shape on the Ni-MMT. The evaluation experiments show the effectiveness of the riblet surface for the positioning accuracy, especially in the case of high speed actuation. In addition, since the MMT is covered by SU-8, which is biocompatible, it can be applied to cell manipulation without harm.

View full abstract

In this paper, a novel local injection system was proposed using a fluorescent coated nanoprobe into Caenorhabditis elegans (C. elegans) using nanomanipulation system inside electron microscope (E-SEM). The nanoprobe was fabricated by focused ion beam (FIB) process and coated by a fluorescent dye. Its size is needed to be small for a minimal damage with a sharp tip and enough rigidity for the nanoinjection. The proposed direct nanoinjection technique is important to transport fluorescent materials or specific biological proteins into its specific cell for In-vivo experiment.

View full abstract

This paper describes a micromanipulation which automatically performs picking-up and releasing of a micro object with an optical microscope and two micromanipulators. The picking-up of a micro object requires an exact distance of depth between a micro object and a tip of the micromanipulator. The authors propose a depth estimation method which is using difference between the blur widths of a micro object and the tip. By using the proposed method, the tip is able to move to the position of a micro object. The releasing of a micro object is performed with a vibration motor. The vibration of the direction of depth is able to release a micro object to under the micromanipulator. In the experiments, the micromanipulation is completed the automatic picking-up and releasing of a micro object. Effectiveness of the micromanipulation is shown by these experimental results.

View full abstract

Micromanipulation techniques contribute to the progress of medical and biological field. We have developed two-fingered microhand which is using the parallel mechanism to realize precise and stable micromanipulation. As previous microhand has small workspace and large size, we aim for development of a two-fingered microhand which has large workspace and compact size. In this paper, we propose a microhand contains a new parallel mechanism. Firstly, we analyze a current parallel mechanism which is similar with designed mechanism. Secondly, we consider the singularity of parallel mechanism. Lastly, we change the order of joints to have compact. We establish an analytic theory for proposed model and then we analyze the workspace. Furthermore, we draw CAD data to realize the system in 3D. Workspace of proposed microhand is larger than previous microhand with respect to simulation results.

View full abstract

In this paper, a novel application of Carbon Nano-Coils (CNCs) as intracellular nano-robots is described. We succeeded in fabricating nano-robots made of CNCs, and disperse them in water by coating 2-methacryloyloxyethyl phosphoryicholine (MPC) polymer. Then, it swam in medium with magnetic field generated by three pairs of magnet coils facing each other. The facing magnet coils generated a uniform magnetic field between themselves, and its magnetic flux density was 5 mT. The obtained CNC had dimensions of 700 nm (diameter), 250 nm (thickness), 350 nm (pitch). They swam by rotating their body in the velocity of 1.74 μm/s, the thrust force of 0.1 pN, and successfully control its direction of movement by switching magnetic field.

View full abstract

In this paper, we discuss a development of air-flow based on-chip single cell dispensing system. To improve success rate of single cell dispensing, the developed system is composed of a cell loader, an inkjet mechanism and two pair of capacitance sensors to dispense a droplet with a single cell. The performance of the sensor and the inkjet are confirmed using a developed platform with a fabricated microfluidic biochip. Based on these characteristics, the disposable microchip is fabricated by photolithography technique. Finally, we succeeded in automatic dispensing of a single bead with the diameter of 100 μm using developed cell dispensing system.

View full abstract

In this paper, we newly propose an on-chip biotope capable of culturing, manipulation, and sensing to figure out functions and mechanical property of microorganisms. To evaluate the applied force to P. laevis by MMT (Magnetically Driven Microtool), the frame structure is attached to the tip of MMT. The design, fabrication, and evaluation of the force sensing structure with a displacement magnification mechanism of the beam deformation are shown. Finally, the basic characteristics of the P. laevis are investigated by using the developed MMT which is placed into the microchip.

View full abstract

We have developed biochip to isolate gastric cancer cells from the peritoneal washing. Gastric cancer cells, which tend to be agglomerated in the peritoneal washing, are bigger than the other cells and the fragment of soft tissue. In this paper, we present a novel biochip to enrich ultra low-concentrated cancer cells in the peritoneal washing by size-dependent cell filtration. This biochip has an advantage to avoid high-accuracy control of the flow rate because the cancer cells were filted by the gap of the microchannel. Two-step exposure was carried out to achieve the precise 3D uneven channel for cell filtration. Furthermore by setting fins in the microchannel to assist the sheath flow, the capability to enrich the cancer cells was improved. As a result, we succeeded in double enrichment of the cancer cells in the quasi peritoneal washing.

View full abstract

Spheroids which cells aggregate and form have far greater functions than individual cells. In particular, hetero-spheroid composed of different cell species such as hepatocytes and endothelial cells express tissue specific function at a high level, which will realize more precise drag screening and biological research. In this study, we propose rapid formation of three-dimensional hetero-spheroid consisting of hepatocytes and endothelial cells using micro-rotation flow. Hepatocytes are covered by collagen gel layers, whose thickness are less than 200 μm, to increase adhesion strength between hepatocytes and endothelial cells. Gel-coated hepatocytes and endothelial cells are collected to form the hetero-spheroids in array by micro-rotational flow in two minutes. This array could control the size of three-dimensional spheroids hydrodynamically with the standard deviations of less than 19 % by varying the cell density of the medium without altering the device geometry.

View full abstract

Research of biochemical analysis system using micro/nano flow channel and fluid device has been ongoing actively. In micro/nano flow channel, effect of surface condition is larger than that of normal scale. Thus, physical or chemical characteristics such as surface roughness, zeta potential largely affect a system. In this paper, we used hydrofluoric acid, plasma dry etching and machining to etch glass surface. We measured the relation between time and etching depth, and successfully indicated the relation between etching depth and surface roughness or zeta potential.

View full abstract

A microfluidic device based on hydrodynamic fractionation with size was applied for separating muscle precursor cells from heterogeneous primary cells of muscle tissues. Although fabrications of engineered muscle with living cells have been developed, quality of prepared cells is serious for regenerative formation of fabricated muscles. We focused the relationship between cell size and quality of regenerative muscles in vitro. To confirm that, primary cells from rat limbs were separated into 4 fractions by cell size using microfluidic device. The mean size of cells in smallest fraction was around 15 microns. From the observation of muscle cell differentiations of each fraction, it was found that the smallest fraction had high regenerative activity of muscle formations. It suggested the potency of cell processing using microfluidic device for tissue engineering in regenerative medicine.

View full abstract

Recently, study of the micro chemical process that uses slug flow has been active. Slug flow is the flow which appears as striped patterns in micro channels. Slug flow increases reaction speed and mixing efficiency of chemical reactions. In this study, The good of this research generation / separation operations of the slug flows, which is not depend on gravity, flow quantity and the viscosity of the fluid. For this purpose we have developed dynamic slug flow generation / separation technique using electromagnetic valves. We devise a new electromagnetic valve and made a design, manufacture and an experiment. As a result of experiment, the width of the pitch of water and oil can be miniaturized to 0.56 mm and 0.29mm respectively. On slug flow separation, we applied to highly viscous liquid separation.

View full abstract

Numerical simulation in aortic arch aneurysm under pulsatile blood flow is carried out to calculate wall shear stress and blood pressure on vessel's wall. At diastole phase, blood flow rate and wall shear stress increase in the model of aneurysm at lesser curvature. In contrast at systole phase, blood pressure and wall shear stress increase in the model of aneurysm at greater curvature and the model of fusiform aneurysm.

View full abstract

We have been developing Bio-chemical IC chip-set made of glass that has three dimensional micro-channel structures in order to make micro chemical device for biotechnology. In this paper, we report improvement of fabrication accuracy by regulating etching process of the photosensitive glass.

View full abstract

This paper discusses a real time measurement of the rigidity of the red blood cell in the microchannel using a high speed vision system. We focus on the passing time and the length of the cell in the narrow channel as indexes of the rigidity of the cell. The most important advantage utilizing a high speed vision system is that we can see and confirm what is happening in the microchannel while a red blood cell is passing. In the algorithm of the measurement, we check whether a red blood cell is recognized correctly in the channel, and save the only successful data.

View full abstract

We report about verification of allergy detection into the micro chemical chip. For the medical treatment regarding an allergy, it is an important task to detect allergens caused allergies from several hundred kind of allergen. The final goal of our research is to develop the allergy detection chip that enables us to detect specific allergens from a few drops of human's blood. As the first step, the antigen-antibody reaction of the allergen on PDMS plate was verified. After that, the PDMS test chip with chamber of lul amount demonstrated the antigen-antibody reaction successfully.

View full abstract