We have developed a vehicle mounted methane detector by means of 1.65-micron LD (Laser Diode) spectroscopy combined with a mobile GIS (Geographic Information System). The detector is used for high-speed survey of gas leakage from underground pipelines.
The detection limit of the detector is about 70 ppb in optical path length of 1 meter with time constant of 1 second. This high sensitivity is achieved by second-harmonic detection of modulation frequency of the LD the center wavelength of which is stabilized on the center of an absorption line of methane.
The position of the vehicle is determined with accuracy of 1-3 meters by a vehicle positioning machine which is connected to a DGPS (Differential GPS, GPS: Global Positioning System) receiver and a fiber-optic gyroscope. Survey tracks and detection points are displayed and recorded on the mobile GIS which consists of information of pipeline network, streets, buildings and so on.

View full abstract

This paper presents two multiwavelength methods to improve the accuracy of a DIAL system for measuring SO₂ in the lower atmosphere: a dual-DIAL method using three or four wavelengths, and a curvefitting method using five wavelengths. By the selection of appropriate wavelengths, these methods can eliminate the effects of ozone and aerosols. Since there is no significant difference in accuracy between the four wavelength dual-DIAL and curvefit methods, the former is advantageous for SO₂ detection in view of the measurement and data processing speeds.

View full abstract

Application of optically measuring techniques was attempted to develop harmonics voltage and current measuring instruments for distribution systems. To develop the harmonics voltage measuring system using a Bi₁₂GeO₂₀ crystal as a sensing Pockels cell for high voltage distribution lines, zinc oxide elements were adopted as a voltage divider. Also, an optical current sensor system was improved on the harmonics current measuring system using a R₃Fe₅O₁₂ film as a Faraday device. Besides, in a practical three-phase distribution systems the developed high voltage harmonics measuring devices using the optical voltage sensor and current sensor were compared with conventional harmonics analyzers for low voltage.
As a result, it was confirmed that the developed high voltage harmonics measuring devices showed successfully good characteristics in practical use.

View full abstract

For the purpose of expanding optical subscriber network, not only optical fiber cables with high reliability and low cost but also cost reduction of construction and maintenance for optical network are required. Recently remote fiber testing system, which automatically tests for all optical fibers using OTDR and optical switch, has been developed. This system can detect fault location in optical network and immediately give a warning to manager of network.
We investigated practical technologies which are employed in each component of the remote fiber testing system, then achieved its high performance, easy operation and cost reduction. This paper reports keytechnologies of optical switch, coupler module, optical code with filter, test equipment, controller and operation software, that are main components used in the optical fiber testing system.

View full abstract

This paper describes a new fast optical polarization control system by a z-propagating LN (LiNbO₃) crystal with four electrodes on the x and y cut surface. In order to stabilize the fluctuating polarization to right circular polarization, the orthogonal voltages Vx and Vy are applied to the corresponding electrodes, which are directly calculated from the Stokes parameters of input light tapped by a 3:1 divider. The size of the LN element is 1.5×1.5×40mm and the electrode width is optimized as 0.7mm, which is enough to minimize the distortion of electric filed distribution. As the experimental results, all insertion loss was 2.9dB including the tapped loss, polarization extinction ratio was 24dB by the polarization state. The cycle of control was 60kHz using a personal computer. The endless control function is also achieved by making the limit of the maximum change of the angle due to the ratio of Vx and Vy.

View full abstract

This paper describes a field measurement of transient state of polarization in actual OPGW (Optical Ground Wire) caused by lightning. The polarization state and the polarization diversity characteristics of coherent communication in the normal condition are also mentioned. As the results, transient fluctuations of polarization state were detected 51 times in two half months. These data were identified as ones due to lightning by comparing the time records with the data of LLS (Lightning Location System) and the records of power line trip. By analyzing the waveform of polarization state, the rise time is 0.04 to 0.2 msec, which are consistent with the time-to-crest of lightning current. The tail of the wave has damping oscillation of 2 to 5kHz in duration of 2 to 5msec. The largest case of polarization shift is 165 degree, which means polarization changes to almost opposite state. By estimating the locations of lightning by LLS, the transients of polarization are caused by the lightning attacks within 60km distant from the observed OPGW.

View full abstract

We have developed an optical frequency synthesizer with a tuning range of 1THz at 1.5μm region, using three laser diodes and an optical frequency comb generator. The accuracy and the stability of the synthesized output are about 100MHz and 6×10^-10, respectively. The fine setting resolution of 2MHz has been achieved by employing a ‘sub-master laser’ which enabled the elimination of inaccessible frequency regions in conventional methods. For the first time, the synthesizer has satisfied our first step target on four basic performances; accuracy, stability, tunable range, and resolution.

View full abstract

The optical current transformer (optical CT) for GIS (Gas insulated switch-gear) systems was developed using optical fiber as the sensor. The linear birefringence was reduced by twisting the newly developed double-coated optical fiber. The optical CT for metering could measure minimum current of 3.8Arms and had rated current of 8000A. Even at fault current of 180kA peak, the optical CT for protection showed the excellent linearity and no saturation effect was observed. The Japanese standard for electricpower instrument current transformer (JEC1201) was successfully satisfied. No deviation of birefringence was observed over temperature from -40°C to 90°C. The Long term stability of the system was kept within ±0.25% over a month operation. This optical CT has thickness of only 50mm and diameter same as that of the GIS tank. Now, no space is required for installation of the CT. The optical CT will contribute to reduce the area of the electric substations.

View full abstract

In this paper we propose two simple and effective identification methods for the optical network management from using novel optical reflection memories.
One method is used Fresnel Reflection Optical Memory (FROM) set up at the connection point or subscriber terminal and Fresnel reflection measument system based on Optical Low Coherence Reflectometory (OLCR) set up at the telephone office for the single star optical network. The Fresnel reflections from slits in the FROM measured by the reflectometer give some information to operators at the telephone office.
Another method is used Bragg Reflection Optical Memory (BROM) installed at each subscriber terminal for the passive branched optical network. In our poposed system, a different Bragg wavelength is assigned to each branched line subscriber and the level of Bragg reflection from the termination.

View full abstract

In this paper we propose a new learning rule for a spatio-temporal neural network model of the primate saccadic system with a distributed feedback control mechanism. In our model the superior colliculus is represented as the distributed network model and it provides a dynamic control signal to a lumped brain stem model (Robinson-Gisbergen model). Distributed feedforward and feedback weights between the deeper layer of the superior colliculus model and the brain stem model are trained using a finite time interval learning rule based on a steepest descent method. Simulations are carried out on a 20-cell model for horizontal saccades using eye position feedback and velocity feedback, respectively. The model makes accurate saccades to all target locations over the range 2 to 15 degrees even if disturbance is added to the burst generator in the brain stem model.

View full abstract

Evolution of optical broadband local networks from near term systems to future higher grade systems is studied. Direct transmission, wavelength division multiplexing (WDM), time division multiplexing (TDM) and photonic TDM (P-TDM) are considered. These four systems are compared from the standpoints of economy and upgradability. Methods for analyzing costs and upgradability investigated in a previous paper[5] are modified to make it possible to include WDM systems in the analysis.
The results of numerical analyses by using representative examples revealed that the TDM system with copper distribution is most economical for lower capacity applications. The photonic TDM system is shown to be most economical for higher capacity later phase applications. When upgrading costs are taken into consideration, however, photonic TDM can be most economical even for lower capacity systems from earlier phase of deployment. The sensitivities of results to cost assumptions are also studied.

View full abstract

In this paper we consider how to design the automatic choosing control of trapezoid type. A nonlinear function within the given system is approximately represented as sum of piecewise linear equations weighted by automatic choosing functions, whose parameters are optimally selected with the aid of a genetic algorithm. As a result a suboptimal feedback controller is synthesized. The proposed controller is applied to a problem of field excitation control of the power system in Tanegashima. Simulation results show that this controller improves the transient stability remarkably well.

View full abstract

Improved Genetic Algorithms (GAs) have been proposed in this paper. We have directed our attention to “selection” and “crossover” in GAs. Novel strategies in selection and crossover are used in the proposed method. Various selecting strategies have been used in the conventional GAs such as Elitism, Tournament, Ranking, Roulette wheel, and Expected value model. These are not always effective, since these refer to only the fitness of each chromosome.
We have developed the following techniques to improve the conventional GAs: “inverse-elitism” as a selecting strategy and variable crossover range as a crossover strategy. In the “inverse-elitism”, an inverse-elite whose gene values are reversed from those in the corresponding elite is produced. This strategy greatly contributes to diversification of chromosomes. As for the variable crossover range, we combine the following crossover techniques effectively: one is that range in crossover is varied from wide to narrow gradually to carry out global search in the beginning and local search in the ending; another is that range in crossover is varied from narrow to wide.
We confirmed validity and superior performance of the proposed method by computer simulations.

View full abstract

Recently, a design of model following control systems (MFCS) for discrete-time multivariable nonminimum phase systems is proposed using an approximate inverse systems. However, the method is not sufficient since the complexity of input-output relation is not considered. Moreover, the controller structures are not well defined. In this paper, we generalize the MFCS using an interactor matrix and clarify the structures of control systems.

View full abstract

In this paper we propose an improved genetic algorithm named Bee System. The concept of the Bee System comes from behavior of bees: first, a bee finds feed and then it notifies the information to the other many bees by dance to work together. In the proposed Bee System, each chromosome tries to find good solution individually. When some chromosome is regarded as superior one, other chromosomes try to find solution around there using multiple populations. Such a procedure is repeated. The Bee System employs some new operations such as concentrated crossover, and Pseudo-Simplex Method. By computer simulations it is confirmed that the Bee System has very good performance.

View full abstract

This work investigates a new approach to detect human face in a cluttered scene. Our method consists of two main search procedures, both using genetic algorithm. The first one is to locate a head inside the scene and the second one is to identify the existence of face within the extracted head area. For this purpose, we have developed two models to he used as a tool to calculate the fitness for each observation in the search procedure: One is a head model which is approached by an ellipse and the other one is a face template which size is adjustable. To find a face, we exploited face salient features such as eyes and nose/mouth on binary image where the features look darker. The procedures work sequentially. The head search is activated first, and after the head area is found, the face detection is activated. The experiment demonstrates the effectiveness of the method. Since our ultimate aini is to provide a good quality of face image for recognition, we also made comparison between face image magnification by linear interpolation with that obtained by zooming camera. The results gives our preference toward that, obtained by zooming camera.

View full abstract

This paper presents a method for robust model matching control for linear continuou-time systems in the presence of unmodeled dynamics. This scheme can robustly control the nominal model in the presence of unmodeled dynamics and achieve the desired model matching simultaneously. Furthermore, the sufficient condition for stabilizing the nominal model in the presence of unmodeled dynamics is derived and the existence of bounds for all signals is proved. Finally, computer simulation results are presented to show the effectiveness of the proposed method.

View full abstract

In this paper we describe a new technique for detecting the front obstacle from the cruising vehicle using stereo images. To calculate the depth from the camera to the obstacle we have to find out the point of coincidence in two images. However, the process takes much time. So, we attempt to eliminate road surface signs in the left image from the right reverse projection image of the road surface to reduce processing time. Next, we decide the square region that contains an extracted object territory. Then the square region is matched to the right image. This technique was applied to the image acquired on the expressway. As a result, the measurement of the distance to the front obstacle becomes possible.

View full abstract

In a heat source system, which is a part of building air conditioning system, numerous sensors are used to measure the temperature of the water. Low precision sensors often lead to measurement errors. In addition, human carelessness usually gives rise to faults during sensor installation: for example, one sensor is installed in mistake for another one, sensors that are used to measure the water temperature are not steeped in the water, etc. Since the installation faults of sensors present themselves in the form of changes of measured values, we propose a qualitative reasoning based sensor installation fault inspection method for sensor fault detection as: First, construct a qualitative model based on the specifications. Next, modify the measured values according to the assumed sensor installation faults. Finally, identify the faults by using stochastic qualitative reasoning. This method is applied to a heat source system to confirm its effectiveness in detecting the sensor installation faults successfully.

View full abstract

First, we propose a more practical method for realizing nonlinear and linear PID controls by a fuzzy inference method. By applying a simplified indirect method of a fuzzy inference method to a few fuzzy control rules of a simple structure, not only nonlinear PID) control based on a PID controller but also linear PID control can be more simply realized than by previous methods. Next, we propose a hybrid control in which fuzzy and PID controls are hybridized for the purpose of trying mutual compensation of both controls. In this proposal, fuzzy control (nonlinear PID control) and PID) control (linear PID control), which are realized by the simplified indirect method of a fuzzy inference method, are combined in parallel and used alternately. The main practical characteristics of this method are as follows: high-speed control can be realized by applying the simplified indirect method of a fuzzy inference method; and the smoothness of switching control can be guaranteed by fuzzy-switching of fuzzy and PID controls.

View full abstract

This paper presents a new design method of optimal continuous deadbeat servo-system by using the second-order smoothing element.
The continuous deadbeat controller is composed of both the series integral compensator and the local feedback compensator introduced into the state feedback loop. The decision method of the damping factor and the natural angular frequency of the smoothing element is described.
A numerical example is given to show how well input-output characteristics are improved, and it is found from a experimental results that the proposed method is satisfactory for practical systems.

View full abstract

We have proposed a paper currency recognition system by a neural network (NN) to develop new type of paper currency recognition machines. This system is constructed by some core techniques. One is the small scale neuro-recognition technique using masks. The other is the mask optimization technique using genetic algorithm (GA). In this mask optimizing technique, I have shown the effectiveness of the adopting GA. Up to now, it has been shown that the NN recognition and learning system are enough as a design and developing method in the market. However, in case of the mask optimizing system, it has not been shown yet. In this paper, I and my group implement this mask optimizing system using GA from EWS to personal computer. Then I show its pragmatic effectiveness and generalization using Italy, France, and Spain currencys. Furthermore, I show the relationship between the least masked area and mask optimized condition by the simulation result. Finally, I also refer running method of this mask optimizing system to aim at using as design and developing tools.

View full abstract

The learning automaton is one of the best known learning systems which is capable of operating in highly uncertain random environments. In this paper, we propose a new learning system which accelerates the learning speed by introducing communications between learning automaton networks. In the proposed system, learning automaton networks can quickly respond to the dynamical changes of the random environment because they can share the information about the random environment by communications. We apply the proposed system to a learning problem of controlling the behavior of artificial insects. Computational experiments show that proposed system obtains better results in terms of the convergence speed comparing conventional learning automaton networks without communications.

View full abstract

In the complex adaptive systems interesting blobal behaviors emerge from many local interactions. When the emergent behavior is a computation, we refer to it as an emergent computation. The premise of the emergent computation is that interesting and useful computational systems can be constructed by exploiting interactions among primitive components.
In this paper, we present a collective model of learning automata which is a general reinforcement learning system such as multi-agent control system. In our model, each automaton has a behavioral tactic directed value of utility function that explicitly depends on the automaton's strategy and the corresponding response from automous fields. The goal of the automaton as well as the nature of their interactions is assumed to satisfy certain conditions which ensure the existence and uniqueness of a Nash strategy. For further studying and taking some applications with this model, two formal examples of pursuit-evasion problem and optimized resource assignment problem are presented.

View full abstract