[in Japanese]

View full abstract

Neutral networks, which are landscapes containing neighboring points of equal fitness, have attracted much research interest in recent years.In this paper, we applied steady-state GAs with different deletion methods to two test functions in order to investigate the relationship between the GA performance and the population diversity.The results show that a population conversed in fitness space moves faster on a neutral nework, but can easily get trapped on neutral networks which do not lead to neutral networks of higher fitness as ruggedness increases.

View full abstract

[in Japanese]

View full abstract

Acquiring environmental information and using it for learning or evolutions are significant issues for an intelligent wheelchair with autonomous, cooperative and collaborative behavior. However, the learning with too many dimensional data derives "curse of dimensionality," so that the learning becomes inefficient. In order to resolve the issues, this study proposes a new technique by which continuous environmental data is converted in arbitrary number of dimension discrete data.

View full abstract

We have been researching and developing ACCoMo (intelligent wheelchair as Autonomous, Cooperative, COllaborative MObile robot) in order tosupport independence of elderly people and disabled people.This paper proposes an behavioral decision method which controlsACCoMo in cooperation with external environment using RFID system.ACCoMo equips a reader-writer of RFID tag by which ACCoMo can exchangeinformation with RFID tags installed into external environment.Results of experiments show ACCoMo can behave suitably in unknownrandom environment making use of the information from RFID system.

View full abstract

The purpose of this study is to develop an operation support system for intelligent wheelchair called ACCoMo. ACCoMo is an electric wheelchair which can autonomously move in environment and ease burdens on users. A touch panel is used as a interface for conveying user's intention to ACCoMo. ACCoMo equips a single-eyed camera, and is controlled using the image processing based on DP matching.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

A group is divided into a partial group and a role assignment is carried out to each partial group. Alone, in spite of having only the simple function, the new function whose intellectual action is attained as a large-scale system can be generated. In this research, the brain function was carried out for a role assignment of this group robot and partial group division to reference, the control system was built, cooperation between robots was tried, and creation of a simulation is tried on the computer.

View full abstract

In this paper, reinforcement learning approach to motion control of 2-link planer underactuated manipulator is described. This manipulator has one passive joint and is difficult to control. The experiments of learning to control this manipulator by RL and human are executed. Using the experimental results, the machine learning algorithm reproducing human abilities is considered.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This paper deals with a trajectory tracking problem of a class of bimodal piecewise affine systems.First, we introduce an error variableas a generalization of the tracking error.A discontinuous function is adopted as an error variableto overcome an inherent difficulty in tracking of piecewise affine systems.Next, the system which governs the error variableis stabilized using a Lyapunov-like function,which yields trajectory tracking of the original system.Finally, a numerical example is given to illustrate theeffectiveness of the proposed method.

View full abstract

Because of the existence of constraints and the combinatorial nature of continuous time and discrete event dynamics, the understanding of hybrid systems is rather limited at present. Few theoretical properties of system can be predicted until now.In this paper, we extend the internal model principle to hybrid systems so as to asymptotically tracking a family of reference signals produced by an external generator.

View full abstract

This paper proposes a new design method of a generalized predictive control in a multirate system. New design parameters are introduced to the multirate GPC law and can be designed independently to sampled outputs. Using the parameters, ripples between the sampled outputs are suppressed independently to the sampled outputs.

View full abstract

[in Japanese]

View full abstract

We present an inverse optimal adaptive control design method for nonlinear systems with unknown parameters.A controller is said to be inverse optimal when it minimizes a meaningful cost functional. To solve the problem, we use pointwise min-norm designs based on control Lyapunov functions. In order to construct adaptive controllers, we develop a systematic procedure using backstepping method.Relating to the resulting controllers, we also present a method achieving locally optimal for nominal systems.

View full abstract

Takagi-Sugeno(T-S) model is one of the most powerful method in describing nonlinear systems. Because T-S model can easily be converted into an LPV system, and the controller design and Lyapunov stability analysis reduces to solving the LMI problem. However, since the constructed T-S Model is generally approximation, how to construct an T-S model with less approximation error is important. In this paper, we propose a model based design method for T-S Model which yields no error for some class of nonlinear systems, and discuss some expansion as well as estimation of error bound for general nonlinear systems.

View full abstract

This paper considers the problem of controlling both the planar position and orientation of an underactuated airship with a reduced number of actuators, whose dynamics are subject to bounded uncertainties. A smooth and time-varying coordinate transformation is utilized to reduce the stabilization problem of the airship to that of a linear time-invariant system. A new robust feedback controller is presented for obtaining global exponential stabilization of the airship in the presence of the plant uncertainties.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Hopping motion has been trying to realize by means of controlling together hopping and attitude with only one linear actuator, and it has already been confirmed by numerical simulation. But experimental verification is still under going because of various disturbances. In this paper, continuous hopping is succeeded by making some minor improvement of robot and control strategy.

View full abstract

[in Japanese]

View full abstract

A Pixel Position Extraction method is a way of rationalization in image processing, extracting the coordinate values of only useful pixels in a captured image. For fast measurement of moving objects, the vision sensor system is constructed with a progressive-scanning CCD camera and a pixel position extraction circuit to employ the method. In this paper, we made it clear that the system can be applied to target tracking from the evaluation experiments. By the results, it is clarified that the system is able to shorten the time of labeling process to 1/10 as compared with a traditional system and accurately detect objects in the case of overlapping and occlusion of objects.

View full abstract

The goal of this study is to seek the accurate three-dimensional self-position of a single CCD camera quickly. A necessary condition of measurement is to take pictures of four marker-lines for four marker-points whose absolute coordinate data are known. The points are acknowledged as intersections of marker-lines. The position of the camera is calculated analytically by solving equations consisting of this observation. Acceptable results were got in indoor experiments with this method using a small model vehicle.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Distance sensors using the triangulation method have been usual in practical fields, systematic studies, however, for the accuracy determined by the optical conditions of the sensor have not been performed. In this paper, the authors prove the serious effects on the accuracy by the optical conditions and the image processing, and will propose the improvement of the image processing.

View full abstract

This paper proposes a simple method of range instrumentation using the standing wave radar. In the range instrumentation by the standing wave, the range will be determined by the distance between the nodes which express the minimum amplitude in the function. This function can be generated when the wave frequency is swept. In this paper, the simple technique is derived in which the distance between nodes is calculated using combination of a low-pass filter and a differentiator.

View full abstract

The vision sensor on mobile robots requires a system that marks the locations of moving objects on an indoor floor for verifying its measurement accuracy. To realize the system that can acquire the positions of moving objects in off time, we have controlled the printing head of a bubble jet printer by using electromagnetic waves carrying video signals by means of a UHF TV Transmitter and a tuner unit. In this paper, we describe the circuit and board designed to control the printing head.

View full abstract

As a key for choosing a walking route, it is important to take into account the flow of other pedestrians as well as how it is congested, because how easily you can walk through depends on both of the conditions. This research proposes a system to measure the distribution and the flow of the pedestrians on the pathways. Our method uses fisheye cameras to cover large area by limited number of sensing units.

View full abstract

A 3D map is more suitable for pedestrian navigation than a 2D map which is used in existing systems, because it allows a walker to check the present position and destination from arbitrary viewpoint easily. Therefore, a simple method for generating 3D maps is required, while they are usually prepared by tiresome work based on the design for an underground mall. This research proposes a motion stereo approach using an omni-directional camera, whose position and posture are acquired by an acceleration sensor and a geomagnetic sensor.

View full abstract

[in Japanese]

View full abstract

This paper describes the measurement technique which enables measurement of a motion of the body at the time of everyday life. For the purpose achievement, it is important to decrease discomfort when measuring. As an example, a measurement technique of the knee joints which combined a measurement method of the angle using the acceleration sensor and a measurement method of myoelectric activity is shown.

View full abstract

In order to encourage the spread of distributed energy systems, an autonomous diagnosis is effective for the supporting technology. The purpose of this study is to develop the autonomous diagnosis algorithm using the fieldbus, which makes it possible to share much information of field devices. The authors constructed a experimental apparatus of a hot water supply process and a measurement and control system with fieldbus devices. Through experiments of the process characteristics, the diagnosis method for the air accumulation to shell of heat exchanger and for the operating point changes, in which the overall heat transfer coefficient is always monitored as the representative process parameter, was developed and its effectiveness was confirmed experimentally.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

We proposed a general prediction method based on the efficient computation and online update of the Singular Value Decomposition (SVD) of historical data. In this paper, an application of our method to route travel time prediction is described. Using real travel time data from short sections (links) on freeway, we evaluated prediction performance of travel time on longer section (route) including the links.

View full abstract

The prediction method of the travel time in urban district using the travel time data of the small section which constitutes a route is discussed. The forecasting model consists past time series data about the travel time of the small section. The travel time and accuracy of a near future of the route are estimated using State Space Representation of the forecasting model. The validity of the method is verified using real data.

View full abstract

[in Japanese]

View full abstract

This paper presents a research and development of a multi-agent flow rotary intersection model using a new Petri-Net like multi-agent flow model by a simple hybrid CAD tool. In addition, to show the utility of this model, some simulation results of the dead lock avoidance by a traffic flow control is shown.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Driving a car around a merging point is a stressful task because a driver is forced to pay attention to many objects and make a decision to change lane. The Optimal Path Generator(OPG), which is an algorithm to calculate desirable maneuver from surrounding environment information, is a hopeful tool to relieve driver's stress. To make the OPG suitable for merging, an auxiliary logic to calculate matching velocity for merging is introduced and combined with OPG.

View full abstract

[in Japanese]

View full abstract