As efficient utilization of computational resources is increasing, evolutionary technology based on the Genetic Algorithm (GA), Genetic Programming (GP), Evolution Strategy (ES) and other Evolutionary Computations (ECs) is making rapid progress, and its social recognition and the need as applied technology are increasing. This is explained by the facts that EC offers higher robustness for knowledge information processing systems, intelligent production and logistics systems, most advanced production scheduling and other various real-world problems compared to the approaches based on conventional theories, and EC ensures flexible applicability and usefulness for any unknown system environment even in a case where accurate mathematical modeling fails in the formulation. In this paper, we provide a comprehensive survey of the current state-of-the-art in the fundamentals and applications of evolutionary technologies.

View full abstract

The covariance matrix adaptation evolution strategy (CMA-ES) is a kind of stochastic optimization such as particle swarm optimization (PSO), and has been shown to have a good performance. However, there are few control applications of the CMA-ES except for only one paper. This paper deals with a PID control problem with constraints on sensitivity and complementary sensitivity functions, and proposes a PID controller tuning method based on the CMA-ES. Numerical examples are given to show the effectiveness of the proposed method in comparison with the recently proposed PSO-based method.

View full abstract

In this paper, we describe a direct design approach using a GA(Genetic Algorithm) to a HSTF(Hierarchical Structure Transversal Filter). The HSTF has a hierarchical building block structure of tapped cascaded low order FIR filters. Since this architecture can reduce the sensitivity of the rounded filter coefficients, the word lengths of the Subfilter and the tapping coefficients can be limited with 6 bit and 7 bit respectively. A direct design approach of the HSTF is not easy because of highly nonlinear problem. So far, we took a two-stage approach. However, this approach cannot give an accurate specification of the filter. We applied a GA into a direct design of the HSTF and could get some preferable results. An design example by GA will be given compared with these of the two-stage design.

View full abstract

In recent years, various types of equipment have become more intelligent. In this research, we propose an intelligent lighting system for providing required illuminance to specified locations, and develop autonomous distributed optimization algorithm which enables advanced lighting control. This system consists of multiple intelligent lighting fixtures, multiple movable illumination sensors and a power meter connected to a network. There is no central unit, therefore, the system has high robustness against the malfunction. We propose a new algorithm which provides rapid convergence to the target illumination by using a correlation of the illuminance and the luminance. We constructed with a dimmer experimental system, that is comprised of 15 fluorescent lights, and several movable illuminance sensors. The verification tests were carried out in the different environments using the proposed control method. We confirmed that the algorithm can provide good performance to environmental change and it is effective to energy saving.

View full abstract

Estimation of Distribution Algorithms (EDAs) are a promising evolutionary computation method. Due to the use of probabilistic models, EDAs can outperform conventional evolutionary computation. In this paper, EDAs are extended to solve reinforcement learning problems which are a framework for autonomous agents. In the reinforcement learning problems, we have to find out better policy of agents such that it yields a large amount of reward for the agents in the future. In general, such policy can be represented by conditional probabilities of agents' actions, given the perceptual inputs. In order to estimate such a conditional probability distribution, Conditional Random Fields (CRFs) by Lafferty (2001) are introduced into EDAs. The reason why CRFs are adopted is that CRFs are able to learn conditional probabilistic distributions from a large amount of input-output data, i.e., episodes in the case of reinforcement learning problems. Computer simulations on Probabilistic Transition Problems and Perceptual Aliasing Maze Problems show the effectiveness of EDA-RL.

View full abstract

A delivery route optimization that improves the efficiency of real time delivery or a distribution network requires to solve several tens to hundreds cities Traveling Salesman Problems (TSP) ⁽¹⁾⁽²⁾ within interactive response time, with expert-level accuracy (less than about 3% of error rate). To meet these requirements, a multi-inner-world Genetic Algorithm (Miw-GA) method is developed. This method combines several types of GA's inner worlds. Each world of this method uses a different type of heuristics such as a 2-opt type mutation world and a block (Nearest Insertion) type mutation world. Comparison based on the results of experiments proved the method is superior to others and our previously proposed method.

View full abstract

We propose a new program evolution method named PORTS (Program Optimization by Random Tree Sampling) which is motivated by the idea of preservation and control of tree fragments in GP (Genetic Programming). We assume that to recombine genetic materials efficiently, tree fragments of any size should be preserved into the next generation. PORTS samples tree fragments and concatenates them by traversing and transitioning between promising trees instead of using subtree crossover and mutation. Because the size of a fragment preserved during a generation update follows a geometric distribution, merits of the method are that it is relatively easy to predict the behavior of tree fragments over time and to control sampling size, by changing a single parameter. From experimental results on RoyalTree, Symbolic Regression and 6-Multiplexer problem, we observed that the performance of PORTS is competitive with Simple GP. Furthermore, the average node size of optimal solutions obtained by PORTS was simple than Simple GP's result.

View full abstract

Survival selections are proposed for a new Differential Evolution (DE) based on the continuous generation model. Many of the conventional DEs have employed the discrete generation model. In the discrete generation model, two populations, namely, old one and new one, are used. Also, the members of the new population are generated from those of the old one. On the other hand, in the continuous generation model, only one population is used and a newborn individual is added to the population immediately. Besides better convergence, the new DE has some advantages. For instance, various survival selections can be easily introduced into the new DE. Therefore, three survival selections depending on the distance between individuals are proposed for the new DE. In order to evaluate the effectiveness of the proposed survival selections, not only the numerical experiment but also the statistical test is conducted on various benchmark problems.

View full abstract

A novel RFID tag with sensing functions of biomedical signals is proposed in this paper. This RFID tag is developed for sensing heart rate, respiration, and body temperature of a human, which is intended for a secure and intensive health care of humans at hospital or at home. The features of this RFID tag are identification of indivisuals by ID call and collation, on-demand measurement of vital signals, and a medium-range(<8m) data communication. The developed prototype RFID tag works at ±1.5V supply voltage, and it was implemented on a PCB attached to a belt. Its performances were confirmed with experiments.

View full abstract

We have analyzed cerebral white matter using magnetic resonance diffusion tensor imaging (MR-DTI) to measure the diffusion anisotropy of water molecules. The goal of this study is the quantitative evaluation of schizophrenia. Diffusion tensor images are acquired for patients with schizophrenia and healthy comparison subjects, group-matched for age, sex, and handedness. Fiber tracking is performed on the superior longitudinal fasciculus for the comparison between the patient and comparison groups. We have analysed and compared the cross-sectional area on the starting coronal plane and the mean and standard deviation of the fractional anisotropy and the apparent diffusion coefficient along fibers in the right and left hemispheres. In the right hemisphere, the cross-sectional areas in patient group are significantly smaller than those in the comparison group. Furthermore, in the comparison group, the cross-sectional areas in the right hemisphere are significantly larger than those in the left hemisphere, whereas there is no significant difference in the patient group. These results suggest that we may evaluate the disruption in white matter integrity in schizophrenic patients quantitatively by comparing the cross-sectional area of the superior longitudinal fasciculus in the right and left hemispheres.

View full abstract

We developed a novel real-time DNA detection system for loop-mediated isothermal amplification (LAMP) method. Our prototype was composed of a thermostatic chamber, a hole slide glass, LED and a web camera. The reaction mixture was injected into the slide glass hole and the LAMP reaction was carried out at 63°C for 2 hours. To observe the DNA amplification, we monitored the fluorescence intensity of SYBR Green I that was excited by the blue LED. The captured BMP images were analyzed by NIH Image J software. The DNA amplification and amplification monitoring experiment was successful. Furthermore, quantitative accuracy was evaluated based on real-time PCR. The reaction time correlates well with the DNA concentration. These results indicate the successful development of a novel real-time DNA detection system for LAMP method.

View full abstract

The pulmonary function test by spirometer is generally conducted. The test subjects, especially children, women and older people, feel uncomfortable as the mouthpiece and nasal plug must be attached to the face of them. We have studied the nonrestraint pulmonary function test using the dot matrix pattern projection in order to decrease the burden to the examinee. In our proposed system, the pattern light projector illuminates the thorax with the dot matrix pattern light. And the CCD camera takes a series of images of the dot matrix pattern. The three dimensional shape of the thorax surface can be calculated by the distribution of light dots. And the respiratory waveform is calculated by the time-series change of the three dimensional shape. The respiratory waveform of our system was similar to one of spirometer. Therefore, we clarified that our proposed system can equivalently measure the respiration with spirometer. And we compared the volume change of the three dimensional shape calculated by our system with the expired tidal volume measured by the expiration gas analyzer. And we examined the relationship between the expired tidal volume and the volume change of the thorax surface.

View full abstract

In the 12-lead electrocardiograph currently being used general medical practice, electrodes are positioned at 6 locations in the chest region and the cardiac potential is measured. This research increases the number of electrode to 124 at evenly-spaced intervals over the body surface of the chest, side and back. The commonly used band elimination filter is not used as a countermeasure for exclusion of the noise from such electrodes, and a body surface electrocardiograph has been developed that makes it possible to perform high-speed sampling of the cardiac potential at 80-100 times the conventional rate. From the sampling data obtained with high spatial resolution, maps and animations of the body surface potential distribution are created and displayed from the 1dimension waveform as well as from the 2dimensions/3dimensions waveforms.

View full abstract

In this study, we purpose a control system of engine torque for a V6 spark ignition engine with a variable valve lift system. We apply a nonlinear receding horizon control to a benchmark problem where we assume that the control inputs are throttle angle, variable valve lift and ignition timing. Moreover, a fuel injection control is also developed by estimating the amount of air in the engine. We propose a vehicle speed tracking control by defining time derivative of input as new input and dealing with input constraint by transformation of input variables. Finally, we show some numerical simulation results to verify the proposed methods.

View full abstract

In this paper stabilization control of nonlinear system of polynomial form is studied using the the strict linearization method. The controlled object is expressed by polynomial form with tensor calculation. The original nonlinear system can be transformed to a linear controllable system by nonlinear coordinate transformation. We can show the global design of stabilization control for polynomial nonlinear system.

View full abstract

Recently, there are various accidents and crimes related to the car. In some cases, the accidents and the crimes can be prevented if it is possible to detect a human who is in the car. For example, we can prevent a baby who is left in a car under the hot weather from dehydration or death occurred by heat inside disease. In another case, it is estimated that the United States currently has as many as 12 million illegal immigrants. In order to prevent further influx of illegal immigrants, the police are physically searching incoming vehicles at national boundaries aiming at finding those who are hiding inside. However, the physical inspections require much manpower cost and time. An inspection method to see inside the vehicles through X-ray images has also been used for this end. But the cost and the installation places are the problems of the large-scale X-ray system. Proposed in this paper is a piezoelectric ceramic system to handily measure the micro vibrations of motor vehicles. And applying the algorithm of Support Vector Machine (SVM), the existence of human body inside vehicles can be detected. The experiment was carried out using four types of vehicles: a mini car; an auto mobile; a van; and a truck weighing 1.5 tons. As the results, the correct determination ratio was 91.2% for the experiment with the piezoelectric ceramic under the front wheels and 97.0% under the rear wheels, when the vehicle used for the examination had also been used together with other three types of vehicles to obtain SVM training data. When the vehicle used for the examination had not been used together with the other three to obtain SVM training data, on the other hand, the correct determination ratio was 93.7% for the experiment with the piezoelectric ceramic under the front wheels and 95.7% under the rear wheels.

View full abstract

In the ubiquitous environment, it becomes important to support people who move in the real world to get information from the Web. In this paper, we consider the situation that the user walking around the real world makes a decision whether or not he/she uses a geographical facility. We develop a system which displays Blog documents about the geographical facility the user got interested in through the mobile terminal. The system searches Blogs on the Web by using not only the facility's name but the keyword about the user's objective. The user just selects the user's objective from the candidates created from common sense database. The system extracts the surrounding text of the selected objective from the Blog documents for decreasing the quantity of the information for the user. The evaluation showed that the system can improve the user's confidence of the decision and decrease the user's effort when he/she has to make a decision in short time.

View full abstract

In this paper, a new framework for removing mixed noise composed of the impulse and Gaussian noises from images is presented in which the FINDRM with the directional difference and the Bivariate Shrinkage Function (BSF) in the Dual-Tree Complex Wavelet Transform (DT-CWT) domain are used. First, the noise detection phase of the Fuzzy Impulse Noise Detection and Reduction Method (FINDRM) is used to determine whether a pixel is an impulse or not. When the pixel is determined as an impulse noise, the FINDRM with the directional difference is used to restore the impulse noise. Second, Gaussian noise is removed by using the BSF, which considers the relationships between wavelet coefficients in the DT-CWT domain. Applying the proposed framework to an image corrupted by mixed noise, a clean image can be obtained.

View full abstract

Lip motion features such as lip width and lip length provide important information to identify individuals or commands. The interface using lip motion features is practical because operation is possible without touching devices. Therefore, it is important to develop a method for extracting the lip shape from the facial image data. In this paper, we propose an automatic lip extraction method on the basis of color and shape information of lips. First, a process already proposed in one of our previous studies extracts lip regions on the basis of color information: difference of redness between lips and skins. Second, lip conditions are distinguished into three type, “Good”, “Protrusion”, and “Defect”, by the difference of the width and height of the lips between object and latest frame. If the lip conditions are “Protrusion” and “Defect”, the correction of lip shape is performed. Our experimental results for eleven persons suggest that the proposed method can extract lip shapes at about 98.1 percent.

View full abstract

A major feature of the self-organizing map (SOM) is a topology-preserving projection from the input layer to the competitive layer, and it has been used mainly as an analytical tool for discovering underlying rules in the given data set. Even though recent splendid progress in this area, there are few novel ideas to break such a conventional style. On the contrary, based on its distinctive nature, a new method for generating patterns through backward projection from the competitive layer to the input layer is proposed recently. Moreover, a promising technology for producing animation as a series of backward-projected patterns along with any pathways on the competitive layer is presented. Then, in order to carry out further considerations, some computer simulations with a variety of posed stick figures are tried in this paper. After training, four kinds of pathways, which correspond to different movements such as dancing, exercising and walking, are prepared. Though some of them does not contain any training samples, all of them worked well as we have intended in advance. As a result, it is found that the proposed method shows good performance and it is also confirmed its effectiveness.

View full abstract

In this paper, we propose a control method for a rotary crane system using neuro-controller (NC) optimized by real-coded genetic algorithm (GA). The rotary crane is known to be a nonholonomic system. We have been successful to suppress the load swing from an initial rotation angle using NC. However, the trained NC have low control performance with untrained angles. In this study, the evaluation function of GA is improved in order to control the load swing from multiple initial positions. The validity of the proposed NC is verified through computer simulation. Simulation results show that the proposed NC has good control performance and robustness with noise and fluctuation of the initial states.

View full abstract

This study focuses on a stock reduction method for build-to-order (BTO) products to flow surplus parts out to the market using sale by recommendation. A sale by recommendation is repeated in an each business negotiation using a recommended configuration selected from the inventory of parts to minimize the stock deficiency or excess at the end of a certain period of the production plan. The method is based on the potential of a customer specification to be replaced by an alternative one if the alternative one is close to the initial customer specification. A recommendation method is proposed that decides the recommended product configuration by balancing the part consumption so that the alternative specification of the configuration is close enough to the initial customer specification for substitutability. The method was evaluated by a simulation using real BTO manufacturing data and the result demonstrates that the unbalance of the consumption of parts inventory is improved.

View full abstract

Two simple differential-type monostable multivibrators are proposed and demonstrated in this letter, where the first configuration operated under positive triggering mode utilizes EX-OR gates and the second configuration operated under negative triggering mode employs EX-NOR gates. Both realization methods make use of CMOS logic gates connected with a few passive elements to form as switching circuits. Experimental results show that the ranges of the output pulse widths are from several microseconds to milliseconds.

View full abstract

Optical signal power estimation utilizing an improved optical fiber identifier is experimentally evaluated. In order to realize fine measurement, a mechanical bend adjuster and fiber guides attached to both side of the identifier are introduced. The former is useful for stabilizing the identifier operation in consequence of well controlled bend of the test fiber. The latter is effective for eliminating influence of fiber handling. As a result, standard distribution of the measured optical power of less than 0.16dB is obtained, which seems to be enough small for the field use.

View full abstract