In early diagnosis of glancoma, intraocular pressure measurement is one of an important method. Non-contact method has measured eye pressure through the deformation of cornea during the increase of the force due to air puff. The deformation is influenced by the cornea stiffness as well as the eye internal pressure. Since the cornea stiffness is unknown in general, it is difficult to evaluate the ture eye pressure. The dynamic behavior of cornea under air puff may provide us with a good hint for evaluating the cornea stiffness appropriately. For this purpose, we develop the sensing system composed of a high speed camera, a mirror for producing a virtual camera, a non-contact tonometer and a slit light source. This system enables us to measure the cornea deformation under concave shape. We show the experimental data for human eyes as well as an artificial eye made by transparent material.

抄録全体を表示

In this paper, we propose a mathematical model for an electromagnet placed inside a molding machine and design a PI controller for this system. The eddy currents inside the electromagnet are a spatially-distributed phenomenon that is difficult to capture by a using finite-dimensional system. First based on our proposed spatially-discretized model, we investigate how this phenomenon makes the system difficult to control and analyze an essential mechanism of the system by using a partial differential equation obtained as a limit of the spatially-discretized model. Second, we propose a new loop shaping procedure that utilizes the GKYP lemma for this system.

抄録全体を表示

It is known that the evolutional computing methods hybridized with local search methods, which are called Memetic Algorithm, are efficient as global optimization methods. The memetic algorithms are divided into two classes of hybridization strategies, which are called Baldwinian and Lamarchian types. This paper is concerned with computational considerations for these two types of hybridization strategies in a case when Particle Swarm Optimization is used as evolutional algorithm from the standpoint of global optimization methods. Especially, applications to optimization problems with constraints described by disconnected plural sets, which are difficult to solve by usual methods, are considered.

抄録全体を表示

In this research, we analyzed the dialogue that consisted of an instruction and a response utterance to clarify the relation between cognitive process and timing structure of utterance. The results showed that correlation coefficient between duration of instruction utterance and switching pause had diverse values when fluctuation of instruction utterance duration was too small to be recognized. However, the correlation coefficient between them had positive and high value when fluctuation of instruction utterance was explicitly big. From these results, we discussed that there are two types of utterance structure in human dialogue.

抄録全体を表示

This paper reports the development of a street performance robot “DKG-MS1” which mimics Japanese traditional juggling “KUWAEBACHI”. In this performance, an entertainer holds a drumstick in his/her mouth and then balances an object like a ball or a teapot on it with great concentration. Regarding this trick as a kind of visual feedback control by human, we create a robot having a rotating arm with a USB camera. To compensate the delay in the visual sensing system and the uncertainties in terms of the balancing object, we design a feedback controller based on the sampled-data H^∞ control theory. The performance of our robot is illustrated with experimental data.

抄録全体を表示

In this paper, it will be clear the structure of the Iterative Learning Control (ILC) based on the inverse system. Moore-Penrose pseudo-inverse of a Toeplitz matrix will be investigated to analyze the learning gain matrix and will be derived the cascade controller transfer function matrix. From these investigations, the critical points of ILC based on the gradient will be issued.

抄録全体を表示

This paper proposes the correlation-based model-free controller synthesis for MIMO systems requiring I/O data acquisition of the same frequency of the input dimension of the controlled plant. Since the controller to be designed is assumed to be a linear regression model, the optimal control parameters are obtained by least squares in the proposed method. Moreover, the proposed method employs the correlation-based cost function, the tuned parameters are insensitive to noises contained in the measured I/O data. The effectiveness of the proposed method is confirmed by some numerical simulations.

抄録全体を表示