It's useful to do the number of people measurement in various situations at a shopping mall. Because it leads to relaxation of the congestion and improvement of the marketing efficiency. But the number of people measurement by visual is difficult. Also, the number of people measurement using the sensor has difficulty in setting place and is expensive. So, I study the number of people measurement using a camera and make a system. I especially suggest technique which considered people of overlap in this study.

Recently, studies have been conducted on applying flying-robot in wide range of areas. We focus studies on indoor guidance systems using flying-robot. For developing them, human tracking becomes an important issue. As a part of human tracking method, this paper discusses estimation of human body orientation using monocular camera mounted on a flying-robot.

A 3d spatial scanning method was proposed to efficiently detect an object at an interested region without the corresponding points search. The method performed a perspective transformation so that the object at the region had no disparity. Coefficients of the perspective transformation were obtained via a calibration stage of a pre-focusing. The object at the region was extracted by evaluating disparities between the pre-focused stereo pair image. The optical flow was introduced to evaluate them. The 3d spatial scanning method was successfully applied to actual stereo pair images taken by an experiment.

A method was proposed for automatically detecting the vacant parking space. The method was based on the pre-focusing method applied to a stereo pair images so that disparity on a virtual plane setting at expected nose position of parking vehicles. The method was characterized by its simplicity and effectiveness. The proposed method was successfully applied to actual stereo pair images captured in a private parking lot.

This paper describes a system of obstacle detection used by improved block matching by Guided image filtering based on in-vehicle stereo camera.This obstacle detection system can be used to finding car and person and other things on road aromatically.Thereby, it provide a method for autonomous driving which applicable to the Eco-drive automation.In addition,it can solve the problem of the traffic accident and energy crisis.In the paper,guided image filtering be used to the real-time obstacle detection system for getting the better result in fog application.This system can transform the data of obstacle to the car controller in real-time and the controller can do something to avoid the traffic accident.This paper also describes the experiment and the result in the driving school.

Color wheel design problem for the DLP projector system is proposed in this paper. A high-resolution image is projected by switching rapidly 2-value images of 24 frames in DLP projector. Improvement of frame rate is achieved by reducing the number of binary images, but the image quality deteriorates. So we consider color wheel design problem to improve image performance using reduced images. We show the numerical example using RGBW color wheel.

In this paper, we attempt to construct a NUI system for TV operation. This system uses a depth camera, and it enables to control TV by combination of gesture recognition and recognition of the number of fingertip. We use a trajectory of right hand position for gesture recognition, and use a depth image of right hand for recognition of the number of fingertip. In an experiment, we evaluate recognition rates of the proposed NUI system.

Since tablet terminal is small and light , there is a feature that carry a simple , anywhere can be used . But there are many complaints about character input. Therefore, we propose an input device that can be used anywhere and keeping portability. In this method first, It adds information of the keyboard by reading the AR marker that has been printed using a camera. At the same time, it detects an area of the fingertip motion by performing a time-series difference. Finally, determining the character input using the coordinate of the detected fingertip and AR marker

In recent years, there are many works with respect to teleoperation system. However, there are still many difficult problems to solve. One significant problem is communication delays between operator and environment through the network. The delay makes control system unstable. This problem was improved by robust control theory.But delay it self is never vaished and affects operability. In general, feedback of information has potential for improve operability. We focus on the visual information using Head Mounted Display. In this paper, We created a Head Rotation Tracking System for teleoperation system, with HMD and pan-tilt-camera. And measure an Influence of delay when look around in Head Rotation Tracking System on traking characteristic of human operator.

In this research, the user interface functions of a remote monitoring system for ocean thermal energy conversion (OTEC) plant using double-stage Rankine cycle are improved. A function to broadly grasp the states of overall OTEC plant using double-stage Rankine cycle is added. A function to simultaneously display multiple graphs of obtained data is also implemented. Furthermore, a function to check the operating state of pumps of OTEC plant is included. The behavior of the improved user interface of the remote monitoring system for OTEC plant using double-stage Rankine cycle is confirmed throgh simulation.

It is important for control education to improve the teaching technique that combines the lecture and experiments using actual equipment. In this study, we developed a education system to help control education based on the model-based development using ball & beam.

We study control systems employ Delta-Sigma modulators as A/D converters. The noise shaping filter in the A/D converter and the observer gains are iteratively designed to mitigate the effect of the quantization error. A design example is provided to validate the proposed design.

In this paper we consider a simplified simultaneous estimation method of states and unknown inputs for mechanical systems. By using an explicit representation of the limiting LQ feedback gain, we propose a reduced order observer for unknown input systems. Effectiveness of the proposed method is confirmed by numerical simulations.

In this paper, we are concerned with a problem of optimization of the linear observations which are used in the stationary Kalman-Bucy filter. Especially, we consider the optimization of the gain matrix in the observation. In the previous works of the authors, the condition of optimality was already obtained by introducing a symmetric matrix called SSNRM which is a quadratic function of the gain matrix weighted by the inverse of the noise covariance matrix. In this paper, a new numerical algorithm to compute the solution is proposed which is applicable to the observations with lesser dimensions than the signal. The results of numerical experiments are provided to show the efficiency of the algorithm.

In this paper, we are concerned with a problem of optimization of the linear observations which are used in the Kalman-Bucy filter. Especially, we consider the optimization of the gain matrix in the observation. In the previous works of the author, the condition of optimality was already obtained for the stationary problem by introducing a symmetric matrix called SSNRM which is a quadratic function of the gain matrix weighted by the inverse of the noise covariance matrix. The same approach is used in this paper to consider the non-stationary case of the problem, and we obtain the condition of optimality. Also, the relation to the solution of the stationary case is discussed.

In this paper, we discuss the stability of stochastic homogeneous systems. Regarding the stability, we show that the homogeneity is strongly related to their convergence rates with respect to homogeneous norms. Based on the stability result, we deal with the stabilization of homogeneous control systems with guaranteed convergence rates. For the purpose of the guaranteed convergence rates, we show a design method of homogeneous stabilizing control laws. We also show numerical examples of the homogeneous stabilization.

This note investigates Lyapunov-type characterization of integral input-to-state stability of nonlinear stochastic systems. The primal purpose is to remove the remote noise orthogonality assumption from a preceding result on the Lyapunov-type characterization. This note focuses on a brief preliminary report on the first result on the removal, and shows how the key development can lead us to a path-wise notion of integral input-to-state stability with an application to interconnected systems. Further results, details and construction of Lyapunov functions for interconnected systems are reported in another paper³⁾.

This paper investigates practical stochastic asymptotic stability of a discrete-time stochastic system. First, we show a sufficient condition for discrete-time practical stochastic asymptotic stability based on stochastic Lyapunov function. Then, we newly introduce a concept of stochastic one-step consistency. Finally, as a main result of this paper, we prove that stability of an approximate discrete-time stochastic system with stochastic one-step consistency implies practical stochastic asymptotic stability of the corresponding exact discrete-time stochastic system. Although an exact discrete-time model of a nonlinear stochastic system is difficult to be constructed in general, this result enables stability analysis of the exact discrete-time stochastic system by only using its approximate model.

In this paper, we deal with discrete-time stochastic systems whose state transition is determined with time-varying random matrices depending on Markov processes. We first consider the case where the Markov processes (i.e., the associated random matrices) underlying systems include no uncertainties, and show a linear-matrix-inequality-based stability condition for the systems, through the use of conditional expectation. Then, we further consider the case where the systems are uncertain in the sense that they are characterized by random polytopes with some deterministic uncertain time-invariant parameters, and extend the result about the above uncertainty-free case toward robust stability analysis.

In this paper, we present a self-triggered control scheme for constrained control systems. The key idea of our approach is to construct a transition system or the graph structure from a collection of polyhedral sets, which are generated based on the notion of set-invariance theory. The inter-event time steps are then determined by a standard graph search algorithm (e.g., Dijkstra algorithm) by minimizing the total cost towards the terminal state. The proposed strategy is illustrated through simulation examples.

In this paper, we deal with large-scaled control systems, in particular, hierarchically distributed structure of the control systems. Our main purpose is clarify the trade-off and the role-sharing of the individual controllers, the upper level global controller and the lower level local controllers. Our idea for this purpose is to introduce a reference model, which represents an aggregated system composed of the local subsystems. Each locally controlled subsystem is designed to satisfy both of the local control performance and the model matching to the reference model. On the other hand, the global controller is designed to satisfy both of the global control performance and the compensation for the model matching errors in the local subsystems. Then we clarify the trade-off by using these quantities from the viewpoint of robust control. Furthermore, we introduce a stochastic property on the model matching errors and discuss the trade-off in a probabilistic sense.

In this paper, a distributed delay-compensated observer for wireless sensor networks is proposed. Each node of the sensor networks aggregates data from the other nodes, and sends the aggregated one to neighbor nodes. All nodes estimate state variables of a system simultaneously. The observer of each node has a form of the delay-compensated observer with multi sensor delays proposed by Watanabe. The effectiveness of the proposed method is verified by a numerical simulation.

We propose a top-down approach for formation control of heterogeneous multi-agent systems, based on the method of eigenstructure assignment. Given the problem of achieving scalable formations on the plane, we present a hierarchical synthesis procedure to compute a state feedback control that assigns desired closed-loop eigenvalues/eigenvectors. This hierarchical control synthesis not only solves the multi-agent formation control problem, but also is computationally efficient. We demonstrate our results by simulation examples.

This paper presents a method for finding algebraic connectivity of a multi-agent system (MAS) with hierarchical network structure having heterogeneous subsystems. The MAS is represented as an interconnection of several kinds of multi-agent subsystems and a main structure, where only one vertex of each subsystem is connected with the structure. The algebraic connectivity of this MAS can be computed by using a corresponding MAS having homogeneous subsystems under a certain condition, which is explored in this paper.

A tailor-made design and manufacturing system based on IoT (Internet Of Things) technology has been developed for making new value co-creation in our project conducted in SIP program. A smart factory system based on CPS (Cyber Physical Systems) for visualizing the proposed manufacturing system has also been constructed. In this paper, the operational methodology of the proposed manufacturing system considering users' demands such as due date and prices is described. The validation of operational methodology is performed using the smart factory system.

The main goal of this research is to make a production schedule considering to add rapid orders in mass customization. For minimizing total due date, combinatorial auction is applied into the planning phase to develop an optimal plan for both companies and customers. In operational phase, a technique for inserting rapid orders at an idle time is proposed using single auction. To evaluate effectiveness of the proposed approaches, computer experiments are conducted.

In this paper, we consider an automatic railway crew scheduling method for a major private railway that includes express trains and local trains. Urban railways have very complex operation schedule. We apply scheduling method based on step by step optimization using partial pairings to avoid combinatorial explosion. In numerical experiments, we evaluate the usefulness by applying this method to actual operation schedule of a major private railway.

The optimization problem regarding optimizing trajectory and sequence is defined as the merging optimization problem. The arrival aircraft merge to land at an airport sequentially while maintaining sufficient time separation. The authors have developed a merging optimization method which is capable of optimizing trajectory and sequence simultaneously and treating nonlinear functions. In this paper, the merging optimization method is modified and included the capability to enable runway allocation, because most large-scale airports include parallel runways. The simulation results show that the proposed method is capable of optimizing also the runway allocation while optimizing the merging of the aircraft.

This paper presents a new mathematical framework for dynamical analysis of a five-axle, threesteering coupled-vehicle system based on multibody dynamics. The framework especially makes it possible to calculate forces acting on the axles of wheels and the rotational joints between a vehicle and a carrier. In the framework, paths of the coupled-vehicle system can be designed so as to avoid the occurrence of extreme internal forces causing the slippage between the wheels and the ground and the destruction of the joints. The validity of this framework has been verified in computer simulations of a motion of climbing a slope.

This paper proposes a slip avoidance method based on model predictive control. In particular, the slip avoidance is achieved by imposing the constraint so that the force acting on each tire does not exceed the friction circle. The vehicle is assumed equiped with rear right and left wheels which can be driven independently and the front wheel which can be steered. The liner tire model is used to design the controller. The objective function of the model predictive control aims tracking to the reference vehicle turning motion and velocity. The effectiveness of the proposed method is examined through simulations using the non-liner tire models to represent the behavior of realistic vehicles.

To control micro electric vehicles with four in-wheel motors, a design method has been proposed. The method divides the vehicle model into the three layers and designs a controller for each layer. For the rigid motion model, which is one of the three layers, the reference model is nonlinear whose yaw rate depends on the vehicle speed. Nevertheless, the method deals with a linear reference model. This paper considers to design a controller by adopting model matching for the nonlinear reference model. In particular, we show that the linear design method can be applied to this nonlinear model matching problem, through the analysis of the closed-loop system designed by the linear method. By confirming that the response of the designed control system matches that of the reference model, it is elucidated that the designed controller is valid.

Recently various kinds of system to control vehicle mobility are developed. Especially a heavy vehicle tends to lose running stability. For this reason, research on its rollover prevention such as Vehicle Dynamics Control system utilizes differential braking has flourished. This paper is focusing on Active Air Suspension control design approach to solve the rollover prevention problem of a heavy vehicle. The objectives of this method are to prevent or reduce the likelihood of rollover and to make the vehicle more closely follow the driver's intended path during going around a sharp curve.

Global warming and the depletion of fossil fuels encourage installing a large scale PV (photovoltaics) system. The operation and management methods for responding to output suppression instruction are needed. This paper investigates the distributed management problem of PCSs (power conditioning systems) which are used to interconnect the PV system to the power grid. We consider the real-time pricing strategy of the utility, a management office of the PV systems, and each PCS determines own reference input to the active power by solving the optimization problem including the provided virtual price. This feedback interaction of the utility and PCSs suppresses amount of active power output at the point of connection. The effectiveness of the proposed distributed management methodology is evaluated through the physical experiments.

We propose a novel real-time pricing method for load frequency control in a grid when there is a lot of wind power. We control batteries and thermal power plants simultaneously adjusting the electricity price for the owners of heat pump water heaters. In our method, the uncertainty due to indirectly controlling electricity consumption by price presentation is compensated by directly commanding batteries and thermal power plants. Our numerical simulation demonstrates that supply and demand are balanced and the load frequency in the grid is kept within the permissible range when wind power generation fluctuates rapidly in a short period.

Occurrence of maximum electricity saving behavior in an apartment complex where residents manually saved electricity in response to critical peak prices was estimated from the experimental results obtained in Kitakyushu Smart Community. Occurrence of the behavior differed between summer and winter peak hours.

This paper considers power supply strategies in futures markets in the presence of day-ahead markets, whose participating companies bid their supply curve and amount of futures transactions given by an iterative learning method. In particular, we analyze the relationship between the supply strategies and the resulting contracts through simulation. The results reveal that generating companies joining futures markets will reduce their bid price in day-ahead markets close to marginal price, but that the only company attending futures markets tends to heighten the corresponding spot price conversely. We also demonstrate the effectiveness of a dynamic learning of contract strategies in futures markets.

Aggregators are expected to play an important role in implementing demand response in the future energy market. In this research, we present a mathematical model of energy demand networks with a utility company, multiple aggregators and many consumers, wherein the demand of the consumers is optimized through prices determined by the aggregators. In addition to determining the price sent to consumers, aggregators also report their own utility functions to the utility from which the energy price is determined by the utility company. In the present model, the aggregators are modeled as strategic agents to maximize their own benefits and market power by appropriately designing the utility functions to be reported to the utility company. We then establish a model of the aggregators’ decision-making. Finally, we analyze the effects of the aggregators’ decisions on the entire demand network using the proposed model.

Convergence of a stochastic consensus is investigated for a multi-agent systems over noisy directed networks with linear symmetric agents. A explicit relation between the number of iterations and the closeness of the agreement is also established under the noisy environment. The results are illustrated through a numerical example.

Sensitivity due to changes of consensus network topologies has been analyzed based on the L₂ gain from the disturbance before the topology change to the control output after the topology change. In this result, the analytical form of the gain can be obtained in a special case. However, in that case, the gain is independent from the topology change. This paper clarifies the mechanism and provides a self-contained proof for the result that the gain is invariant against the topology change.

This paper deals with distributed power demand-supply management based on dynamic electricity pricing. The purpose of this paper is to maximize the social welfare and to keep demand supply balance. In this paper, consumer, generator and Independent System Operator (ISO) participate in the electricity market, and each consumer and generator act based on regret matching which is one of the learning algorithms. In this paper, we propose a novel procedure to decide the electricity price based on a gradient method with the constraint of AC power grid model. In addition, we discuss the convergence of the proposed algorithm. Finally, simulation results show the effectiveness of the proposed algorithm.

This paper introduces the usefulness of RBF-based soft sensor for melt index of polymer. From examining the findings, RBF-based models brings smaller estimated errors than linear regression models. Here we also discuss the comparison of 3 kinds of RBF-based soft sensor.

Design of experiment performed effective automatic measurement, reduction of experiment point and search of most suitable point, while meeting precision prescribed of exhaust performance and fuel consumption performance of engine. However, conventional limit area search method to use by automatic measurement coordinates plural control parameter in an equal distance. Therefore control parameter having high limit sensibility becomes limitation, and automatic measurement area becomes smaller than original drive limit area. This study suggests limit area search method that variable distance coordinates value of plural control parameter depending on limit sensibility. Precision prescribed of exhaust performance and fuel consumption performance of engine in this way improves.

This paper presents a two DOF controller for a liquid container by an overhead traveling crane. The feedforward part of two DOF controller is based on a notch filter which is designed to cut off a frequency of a liquid vibration. The feedback part of two DOF controller is designed using H₂/H_∞ controller design method. Experiments are performed to show the effectiveness of the proposed controller.

Tokyo Gas actively used dynamic plant simulator in LNG terminal construction. Operator Training System (OTS) was constructed before the startup of the plant not only to apply it to DCS software review, but also to use in making preparations for the commissioning of the plant. Dynamic simulator linked to DCS was proved to be useful in supporting operators by forecasting the process behavior of the plant.

In this paper the memory-based dominance is implemented in the branch-and-bound algorithm for the block relocation problem. In the memory-based dominance technique, nodes in the search tree are compared with those already traversed in order to eliminate dominated and thus unnecessary nodes. Since a node is associated with a block configuration in the branch-and-bound algorithm for the block relocation problem, a new representation of a block configuration is proposed not only to store a large number of traversed nodes efficiently but also to check the dominance easily. The effectiveness of the proposed method is examined by numerical experiments.

In this paper, we study the extension and application of first-order hybrid Petri nets (FOHPN) to the simultaneous optimization problem of traffic signal control and route selection. We propose a decomposition algorithm for traffic signal control and route selection problem via first-order hybrid Petri nets. The traffic signal control problem is represented by the optimal firing sequence problem for first-order hybrid Petri nets. The Lagrangian relaxation method is applied to solve large scale problems efficiently. Computational results show the effectiveness of the proposed method.

The introduction of energy management systems (EMSs) makes it possible that a consumer reduces an electric power purchase cost by power consumption control and operation planning of equipment based on the variable electricity prices. Because the prices in the electric power exchange (EPX) are uncertain, a planning method considering the uncertainty is needed. We have been studied a distributed EMS (DEMS) in which trade of excess energy between EMSs is possible. In this paper, we consider the case where the electric power is supplied from EPX to the DEMS and study a planning method using stochastic optimization.

In this paper, we deal with large-scaled optimization problems. Motivated by the reduction of data transmission and computation in distributed optimization algorithms and the necessity of a criterion for clustering agents in multi-agent systems such as electric power network, we clarify the relationship between the optimal solutions of the original large-scaled optimization problem and the corresponding approximated optimization problem. Based on this analysis, we propose a hierarchically distributed optimization algorithm which reduces the data transmission and computation considerably. Furthermore, we assume stochastic dispersion on the profit functions of the agents and analyze the stochastic characteristics of the solution of the optimization problem.

This paper is concerned with the formation control problem of wheeled vehicles moving on a slope. Due to the non-holonomic constraint, the gravitational disturbance acting on the vehicles changes depending on their heading angles. To cope with this difficulty, we propose a new formation control method with the aid of disuturbance observers which estimate the gravitational disturbances.