This article has no abstract.

View full abstract

We proposed a new architecture for a phase-locked loop (PLL) obtained by comparing clock periods. We evaluated the use of a clock-period comparator (CPC) for the digitally controlled PLL we propose, where only the frequency should be locked. However, frequency control with the CPC resulted in the phase being locked. Thus, phase-lock operation was also achieved. The theoretical analysis of the phase-lock mechanism was confirmed through system simulations. We discussed about dead-zone problem caused by a time delay of circuits. We evaluated phase-shift direction detector to solve the dead zone problem. We designed the element blocks of the new PLL using a 0.25-μm CMOS process. We confirmed phase-lock operation through SPICE simulations of the MOSFET level. Moreover, we manufactured a trial circuit for the new PLL. We also confirmed phase-lock operation in the proposed PLL through measurements.

View full abstract

We propose a forced chaos generator with a CMOS variable active inductor circuit. The equivalent inductance of the variable active inductor in the proposed circuit can be controlled by an external voltage. Therefore, the oscillation frequencies of the circuit can be altered by applying an external periodic square waveform. As a result, we can generate chaos from the circuit. We then confirm the folding-and-stretching mechanism of the chaotic motion in the circuit. Complex phenomena, observed in the proposed circuit, are analyzed through the Poincaré sections from the SPICE simulations with TSMC 0.35μm CMOS semiconductor process parameters. In addition, we define a return map on the Poincaré section to examine the properties of the observed attractors. Moreover, we investigate the bifurcation phenomena when the amplitude and period of the external signal are changed as bifurcation parameters.

View full abstract

This paper describes a new techniques of reducing phase noise in oscillator circuits. Our method uses an external crystal resonant circuit that acts as a frequency reference and is based on correlation with negative feedback control. We present the circuit configuration and the transfer function used in this method, as well as measured single sideband (SSB) phase noise characteristics. Our experiments show that phase noise can be decreased as it is a theoretical value when using LC oscillator. Furthermore, we examine application for voltage controlled crystal oscillator (VCXO). As a results, we can improve that the phase noise characteristics more than that of original VCXO without spoiling frequency tuning range of VCXO.

View full abstract

The Hodgkin-Huxley (HH) equations of a squid giant axon are the most important mathematical model in electrophysiology and biology, and are also important in the development of various biologically inspired intelligent devices such as artificial neuron device and neural networks. The HH equations, however, include various constants or parameters whose values were determined based on physiological experiments, and thus the values possess inherent ambiguities. Also, the ‘constants’ are not really constant but change temporally. Thus, in this paper, we study the effects of the change of the constants or parameters on the dynamics of the HH equations and consider the parameter dependence and sensitivity of the equations; we study the bifurcation structure of the HH equations by changing their various parameters. In particular, we take the voltage-dependency of the dynamics of so-called gating variables as typical bifurcation parameters and show that the HH dynamics is very sensitive to the steady-state functions but not sensitive to the time ‘constants’ of gating variables.

View full abstract

In our previous study, we considered synchronization in a system of two unidirectionally coupled neurons with a chemical synapse. We analyzed the inhibitory and excitatory couplings, assuming that the firing frequency of a pre-synaptic neuron is varied. We found that the neurons easily acquire synchronization at a firing frequency higher or lower than the fundamental frequency of a post-synaptic neuron for excitatory or inhibitory coupling, respectively. In this study, we add one neuron to the previous model; the added neuron drives the previous two neurons (sub-system) by synaptic inputs. To avoid model dependence, we used the Morris-Lecar and Erisir's fast-spiking neuron models as single neuron dynamics. We found that the neurons achieved synchronization at a firing frequency lower (higher) than the synchronous firing frequency of the sub-system for the inhibitory (excitatory) synaptic connection.

View full abstract

A number of studies have recently been made on multi-stable circuits. We previously proposed a multi-stable circuit using Λ-shaped negative resistance devices constituted by enhancement-mode MOSFETs and voltage sources. However, because this circuit contains a lot of biasing circuits, it is disadvantageous in points of the area and current consumption. Our research, aim at construction of artificial neural network equipped for the learning function using plastic synapses which are multiple-valued memory cells. In this paper, we propose the circuit configuration into which cut down the components in the bias circuit of the voltage, and we study the low current consumption of proposed circuit. As a result, it is shown that the bias circuit cut down the number of 2(n-1) (n: number of multiple-valued) MOSFETs. Furthermore, it is shown that proposed circuit works [nA] unit, because the bias circuit connects the state of the high impedance to the driver circuit.

View full abstract

Switched-capacitor (SC) circuit has been a well known analog circuit block. Since their characteristics are determined by capacitance ratio, SC circuits are suitable for on-chip implementations. Eventhough there are so many publications on the design of SC circuits, only a few of them discuss the property of noise in SC circuits. Linear noise analysis are common in continuous-time systems. However, SC circuits are operating in discrete-time domain and their noise properties are known to be a non-linear phenomenon and dificult to be analyzed by hand. This paper will show a noise analysis of an SC integrator as an example for a better understanding of the analysis method. It will be shown by simulation results that the proposed formula gives a better approximation of noise power spectral density (PSD).

View full abstract

In this study, we propose a Spice-oriented algorithm for peak search and stability assessment for the frequency response. We obtain the frequency response by using the sine-cosine circuit, Fourier transformation circuit and solution tracing-curve circuit, which have been developed in our previous studies. We combine the idea using a differentiator and a nonlinear limiter with them in order to search the exact peaks and propose a stability assessment algorithm based on the Floquet theory. In our algorithm, we can simulate almost all tasks with Spice.

View full abstract

In this paper, we propose a novel neural network considering deep cases. It can learn knowledge from natural language documents and can perform recall and inference. Various techniques of natural language processing using Neural Network have been proposed. However, natural language sentences used in these techniques consist of about a few words, and they cannot handle complicated sentences. In order to solve these problems, the proposed network divides natural language sentences into a sentence layer, a knowledge layer, ten kinds of deep case layers and a dictionary layer. It can learn the relations among sentences and among words by dividing sentences. The advantages of the method are as follows: (1) ability to handle complicated sentences; (2) ability to restructure sentences; (3) usage of the conceptual dictionary, Goi-Taikei, as the long term memory in a brain. Two kinds of experiments were carried out by using goo dictionary and Wikipedia as knowledge sources. Superior performance of the proposed neural network has been confirmed.

View full abstract

We propose a nonlinear filter that is effective in removing both Gaussian and impulse noise. The proposed filter achieves an excellent performance in the restoration of images degraded by mixed noise, combining smoothing for Gaussian noise removal and inpainting for impulse noise removal on the basis of total variation minimization principle. We also propose how to determine smoothing parameter λ that is important for the quality of restored images, according to a surface model of λ with respect to σ_n and σ_s, standard deviations of Gaussian noise and original image respectively. In applying the proposed method to images degraded by various mixed noise, we verified that both PSNR and subjective image quality are significantly improved, compared with the conventional filters.

View full abstract

Coupled Markov random field (MRF) models have been proposed so far for visual image processing. These are classified into boundary- and region-based models, where hidden variables that determine the interaction between the units corresponding to image pixels are given by line and label processes, respectively. In this paper, we have investigated a region-based coupled MRF model with phase variables as the hidden variables, and have modified the model for applications to coarse image-region segmentation by replacing some nonlinear functions in the update equations of the intensity and label processes with piecewise linear (PWL) functions. Using PWL functions facilitates VLSI implementation of coupled MRF models, and also makes their performance improved. We have verified that the modified region-based MRF model is superior to the resistive-fuse network, which is one of the boundary-based MRF models. We also propose an improvement of the modified region-based MRF model by introducing a new parameter, and evaluate the model performance in coarse image region segmentation.

View full abstract

An image processing system for beautifying human face images is proposed using interactive evolutionary computing. Authors proposed before a nonlinear digital filter bank system, which removes undesirable skin components from human face images to make the face look beautified. However, it was difficult to set the parameters. In this paper, interactive evolutionary computing is newly introduced to optimize the system, considering the user's subjective criteria and taste. Moreover, edge enhancement is involved in the system to make the face look sharp and clear. The proposed system is verified to be effective in processing an actual face image and subjective test for some people.

View full abstract

This paper describes mixed constrained image filter design for noise reduction using a Genetic Algorithm (GA) with parameterized uniform crossover. The proposed method with GA autonomously synthesizes a filter suitable for the reconfigurable processing array, evaluating the complexity, power and signal delay in both Configurable Logic Blocks (CLBs) and wires. An image filter for noise reduction is experimentally synthesized to verify the validity of the proposed method. By the evolution, the quality of the optimized image filter is better than that reported in other papers.

View full abstract

The tabu search was implemented on a neural network with chaotic neuro-dynamics. This chaotic exponential tabu search shows great performance in solving quadratic assignment problems (QAPs). To exploit inherent parallel processing abilities of analog hardware systems, a synchronous updating scheme, where all the neurons in the network are updated at the same time, was proposed. However, several neurons may fire simultaneously with the synchronous updating. As a result, we cannot determine only one candidate for the 2-opt exchange from the many fired neurons. To solve this problem, several neuron selection methods, which select one specific neuron among the fired neurons, were proposed. These neuron selection methods improved the performance of the synchronous updating scheme. In this paper, we analyze the dynamics of the chaotic neural network with the neuron selection methods by means of the spatial and temporal mutual information. Through the analyses, the network solution search dynamics of the exponential chaotic tabu search with different neuron selection methods are evaluated.

View full abstract

The novel CMOS ring oscillator array focused attention on the adjacent relationship between the component oscillators is proposed. Compared with the conventional CMOS ring oscillator array, the proposed oscillator array has the bi-directional connections. To verify the operation of the proposed oscillator, using 1.2μmCMOS Level 3 models, the circuit simulations are performed.

View full abstract

In this research we physically built several robotics modules that are able to self-discover a connection topology which allows them to generate a coordinated behavior as an integrated modular robot. We consider that this self-configurability of hardware module can potentially simplify the costly designing process of complicated robots and at the same time improve the resiliency of modular robots in the face of internal and external changes.

View full abstract

We have developed a discrete-time dynamical system for dynamic image segmentation. It consists of a global inhibitor and modified chaotic neurons that can generate oscillatory responses. Dynamic image segmentation is performed using its oscillatory responses. This letter presents an implementation of our system in a field programmable gate array (FPGA) device and a successful result of dynamic image segmentation.

View full abstract

Transient electromagnetic fields caused by electrostatic discharges (ESDs) are known to cause a serious failure in high-tech information devices. From this perspective, an international standard for ESD testing or IEC 61000-4-2 has been specified, in which a detailed waveform of the discharge current to be injected onto equipment under test (EUT) by an ESD-gun is prescribed, and is also required for contact discharges onto a calibration current target called Pellegrini target, which is specially designed so that the waveform of an injected current and its observed voltage coincides. However, the coincidence relationship should not hold at high frequencies due to parasitic components which the target has. In this study, we constructed an FDTD model of the Pellegrini target, and calculated its transfer impedance by simulating the injected current onto the target and the corresponding output voltage in combination with an FDTD model being previously developed for an ESD-gun. As a result, we found that the target has a transfer impedance of almost 1-j0 Ω at less than a few hundred MHz, though it has resonance phenomena around 2-3 GHz, which was validated by measuring scattering parameters of the target through a transmission line with a network analyzer. Using the FDTD-calculated transfer impedance, we reconstructed injected current waveforms for contact and air discharges of an ESD-gun from their observed voltages to reveal that the injected current waveform is almost identical to the observed voltage waveform for the contact discharge, while the injected currents for the air discharges have smaller peaks and longer rise times than the observed voltages do due to the presence of resonance properties of the transfer impedance.

View full abstract

We have developed a practical experimental method to mass-produce and maintain a variation of minimal neuronal networks (“micro neuronal networks”) consisted of a single to several neurons in culture using spray-patterning technique. In this paper, we could maintain the micro-cultures for one month or more by adding conditioned medium and carried out optical recording of spontaneous activity in micro neuronal networks and considered the interactions between them. To determine the interactions between micro neuronal networks, fluorescence changes in several small networks were simultaneously measured using calcium indicator dye fluo-4 AM, and time-series analysis was carried out using surrogate arrangements. By using the spray-patterning method, a large number of cell-adhesive micro regions were formed. Neurons extended neurites along the edge of the cell-adhesive micro regions and form micro neuronal networks. In part of micro regions, some neurite was protruded from the region, and thus micro neuronal networks were connected with synapses. In these networks, a single neuron-induced network activity was observed. On the other hand, even in morphologically non-connected micro neuronal networks, synchronous oscillations between micro neuronal networks were observed. Our micro-patterning methods and results provide the possibility that synchronous activity is occurred between morphologically non-connected neuronal networks. This suggest that the humoral factor is also a important component for network-wide dynamics.

View full abstract

In the area of Intelligent Transport Systems, the multiobjective route selection problem (mRSP) becomes an important key problem that needs to be solved in car navigation system (CNS). In this paper, we propose an effective route selection approach for solving mRSP while minimizing driving distance, driving time and driving cost simultaneously. A new multiobjective genetic algorithm (moGA) called distance Pareto Genetic Algorithm (dpGA) is presented to effectively solve mRSP. The mechanism of the proposed dpGA guarantees good convergence toward the Pareto-optimal front and gives sufficient emphasis on the diversity feature. The fitness function used in dpGA is based on two kinds of distance values, i.e. Pareto distance and crowding distance. Finally, we demonstrate the applicability and evaluate the efficiency of the proposed solution approach by using numerical experiments with the real digital road map data. The experimental results show the effectiveness of the proposed solution approach.

View full abstract

Particle Swarm Optimization (PSO), which has attracted attention as a global optimization method in recent years, has a drawback in that sustainable search cannot be performed until the end of computation due to its strong convergence trend. In this paper, in order to realize a sustainable search in PSO, the improved PSO using concepts of particle ages and digenesis is proposed. In the new PSO, parameters in the update formula are degenerated and a stagnant particle is erased if it loses activity, and then a new search point in which large parameter values are assigned. In addition, information regarding the elite point of all searching points until the current time is reflected to new points in next generation. The effectiveness of the improved method is confirmed through applications to benchmark problems.

View full abstract

Various kinds of sophisticated systems are utilized in our daily life, therefore these systems must have the ability to work considering behaviors of humans around them. In order to realize such a system, a kind of human behavior modeling method is required. In this paper, we propose a new modeling method of human behaviors based on the causality between the situation around a person and the human behavior. In this method, we suppose that a person changes his behavior according to the change of the situation around him. In this paper, a human behavior is discribed by a discrete event and the situation around the person is described by multidimensional time series signal. The causality between the time series data of the situation and the discrete event of the human behavior is defined as behavior rule. In this rule, the multidimensional time series data is modeled by HMM(Hidden Markov Model) in order to consider its temporal and spatial redundancy. Various kinds of human behavior rules are acquired based on the combination between the situation around a person and his behavior. Therefore, a set of these human behavior rules is called as a human behavior model in this paper. In the experiment, the controller operation of a radio controlled vehicle is modeled by the proposed method. The usefulness of the proposed method is examined through some experimental results with the estimation system of the next human behavior based on the generated behavior model.

View full abstract

In this paper, the stochastic Nash games for stochastic Markov jump systems are investigated. First, the Nash equilibrium is defined and the existence conditions are formulated. In particular, necessary and sufficient conditions for the existence of two cross-coupled stochastic algebraic Riccati equations (CSAREs) are developed. Moreover, in order to obtain the required solutions, a numerical algorithm that is based on the Newton's method is proposed. As an important implication, weakly-coupled large-scale systems are also adopted. After establishing the asymptotic structure of the solutions for the CSAREs, the quadratic convergence result that is based on the Newton's method is shown. Finally, a numerical example is given to demonstrate the availability of the proposed method.

View full abstract

A knowledge-based approach to imitation learning of motion generation for humanoid robots and an imitative motion generation system based on motion knowledge learning and modification are described. The system has three parts: recognizing, learning, and modifying parts. The first part recognizes an instructed motion distinguishing it from the motion knowledge database by the continuous hidden markov model.
When the motion is recognized as being unfamiliar, the second part learns it using locally weighted regression and acquires a knowledge of the motion.
When a robot recognizes the instructed motion as familiar or judges that its acquired knowledge is applicable to the motion generation, the third part imitates the instructed motion by modifying a learned motion. This paper reports some performance results: the motion imitation of several radio gymnastics motions.

View full abstract

We have developed the communication system synchronized with video scene which is a topic duration defined according to its content. This paper evaluates whether we can understand the content of video more deeply or not by comments registered on this communication system. Two kinds of user tests using TV program videos are performed. One is the comparison of the communication system synchronized with semantic scene whose average duration is 72 seconds, short scene of five seconds and more shorter scene of each video frame. The other is the comparison of the communication system synchronized with viewing videos and independent of viewing videos.
Experimental results show that the communication system synchronized with semantic scene can collect more comments which are related to video and include opinions and proposals based on experience and consideration.

View full abstract

In current manufacturing systems, production processes and management are involved in many unexpected events and new requirements emerging constantly. This dynamic environment implies that operation rescheduling is usually indispensable. A wide variety of procedures and heuristics has been developed to improve the quality of rescheduling. However, most proposed approaches are derived usually with respect to simplified assumptions. As a consequence, these approaches might be inconsistent with the actual requirements in a real production environment, i.e., they are often unsuitable and inflexible to respond efficiently to the frequent changes. In this paper, a multi-objective job shop rescheduling problem (moJSRP) is formulated to improve the practical application of rescheduling. To solve the moJSRP model, an evolutionary algorithm is designed, in which a random key-based representation and interactive adaptive-weight (i-awEA) fitness assignment are embedded. To verify the effectiveness, the proposed algorithm has been compared with other apporaches and benchmarks on the robustness of moJRP optimziation. The comparison results show that iAWGA-A is better than weighted fitness method in terms of effectiveness and stability. Simlarly, iAWGA-A also outperforms other well stability approachessuch as non-dominated sorting genetic algorithm (NSGA-II) and strength Pareto evolutionary algorithm2 (SPEA2).

View full abstract

In this paper, we describe a combination of the particle swarm optimization (PSO) and the simultaneous perturbation optimization method. Moreover, FPGA implementation of this method is considered. Details of the hardware system are also explained.
PSO is an interesting optimization technique with powerful global search capability. PSO has plural search points as candidates of optima. However, this method does not utilize local information of an objective function such as gradient.
On the other hand, the simultaneous perturbation method is a recursive optimization method using a kind of finite difference. The method estimates gradient without direct calculation. Thus, we propose a combined method of PSO and the simultaneous perturbation method. This method has global search capability and local search one at the same time.

View full abstract

The bid-rigging on public works projects has recently become social problem in Japan. We can regard this issue as a framework of spatial game through simplifying the strategy of tendering. The main purpose of this study is to represent the tendering with multiple bidders as repeatedly playing games only with the opponents of relationships. We can regard cooperation to occur by performing the repeated games as bid-rigging. To examine the characteristics of this cooperation will lead to the discovery of new knowledge that prevents collusion.
From the point of view, in this research by the use of the agent based simulation, we have simulated the tendering recently conducted regarding the disposition of public properties and studied the characteristics of cooperation based on the decision with the suppression of desire. When every group mutually makes the moderate decision, we observe cooperation at high level in the spatial game with the extended strategy expression. On the other hand, it is revealed that the probabilistic change in the connectivity of each group greatly influences that cooperation. This perturbation, therefore, is effective for the deterrence of collusive tendering. In addition, altering multiple parameters, we also show further inspections regarding the obtained results.

View full abstract

Sign sounds are used in various daily situations. However, they are not made by reflecting user's own preference. This study proposed Interactive Genetic Algorithm (IGA) method creating various sign sounds suited with each user's preference and objective. The proposed method created sign sounds by using all musical notes, while a previous IGA method have created different sign sounds by using musical rules for restriction of use of musical notes. To investigate the efficacy of the proposed method, two listening experiments were performed with a concrete system based on the proposed method. Result of the listening experiments creating bright and warn sign sounds showed significant increases of fitness value.

View full abstract