Possible choices of the assistive device for walking difficulties and/or disabilities nowadays are the electrically powered wheelchair (EPW) and the so-called “senior-car”. An EPW generally has a joystick as its controlling gear, which makes it hard especially for the aged to manage. The senior-car on the other hand has a very familiar driving interface, but its utilization might be restricted only to a door-to-door transportation because of its large body. In this paper, we will develop a new compact EPW which can be handled easily by anyone who has a driving experience. The proposed EPW has a steering wheel and a foot pedal as its controls, but does not contain a mechanical steering system nor an engine. All car-like behavior will be synthesized by the two driving motors and an implemented MRAS based controller whose reference model simulates the car dynamics. The force feedback steering wheel emulating steering reaction force will also be incorporated into the proposed EPW to imitate car driving feeling during manipulation and to acquire the handling stability and safety of the EPW. Finally, results of the experiments will be reported to show the validity of the suggested EPW.
This paper deals with multiobjective control problems for systems whose partial states can be generated by a stable system. For given problems we propose a method to synthesize controllers which explicitly contain the stable system. The synthesis method is proposed as a consequence of the following results : (a) there exists an equivalent problem where partial states of a given generalized plant are exactly available; (b) the resultant controller is constructed by the stable system and another controller for the equivalent problem. The proposing method is advantageous owing to the following reasons : (1) it is possible to synthesize, based on LMIs, controllers which explicitly contain stable systems; those stable systems generate partial states of generalized plants; (2) computational complexity in LMI synthesis can be reduced; (3) computational time to produce control input can be shorten. Moreover, numerical experiments clarify the effectiveness of the proposed method in the case of discrete input delay systems.
This paper proposes an initialization of bipolar networks with a hidden layer trained for pattern classification problems. Weights of the hidden layer are initialized so that hyperplanes should pass through the center of input pattern set, and those of the output layer are initialized to zero. The initialization principle for the hidden layer can be equivalently realized by the usual random initialization when the training data are normalized so that their average should be zero. Thus the procedure for our initialization becomes quite simple. From several simulation results, it is confirmed that the proposed initialization shows better convergence than the usual initialization in which all the weights take the random numbers.
In this paper, we propose a new complex-type wavelet which we call the RI-Spline wavelet. We show that the RI-Spline wavelet satisfies the admissibility condition and bi-orthogonal condition. That is, the RI-Spline wavelet can be used for continuous and discrete wavelet transform as mother wavelet. We confirm the effectiveness of this wavelet in continuous and discrete cases by using a model signal and applying it to unsteady turbulence analysis. The main results obtained can be summarized as follows : the eddy structure and the energy transportation process of the turbulence can be shown very well. The unsteady turbulence intensity, which is derived from the discrete wavelet transform, is also effective for examining quantitatively the change of the unsteady turbulence energy in terms of time. Our results demonstrate the advantages of our approach.
In this paper we consider a simple order delivery system. Customers order items at their home to a shop or a restaurant. The shop produces items and one delivery-man delivers them to their home. Each shop retains one finished item until the delivery-man returns to the shop. We analyze the waiting time of customers from the order to the delivery of the items at their home.
“Rinbo” and “Toumo” are techniques of calligraphy. “Rinbo” is a technique of penmanship in which a copy is made by putting the original on the side of a writer. “Toumo” is a kind of techniques for reproduction which is supposed as a tracing technique in which a copy is produced by putting the original under a paper. Connoisseurs have distinguished autographs from copies depending on their experience, knowledge and subjective sence. We deal with three ancient works written by writing brush in China, “Shinsousenjimon”, “Souranjo” and “Chinbimei shinshiken shikoudaiji” as the subjects of this study. It has been supposed that “Shinsousenjimon” was written by “Toumo” However, Prof. Uozumi who is an experienced calligraphist hypothesizes that “Shinsousenjimon” is an autograph or a work by “Rinbo”. It is no doubt from past literatures that “Souranjo” was written by “Toumo” and that “Chinbimei shinshiken shikoudaiji” is an autograph. We examine the density distribution of characters in these images. First, we perform multistage binalization of characters in order to compare overall density distribution in respective characters. Next, we measure the distances between most dark lines and geometrical central lines in several straight parts of characters. Lastly, we compare the density between symmetrical points in several lines of characters. As the result of these analyses, we found that “Shinsousenjimon” has characteristics in more common with those of “Chinbimei shinshiken shikoudaiji” than those of “Souranjo”.
Recently models of neural networks that can deal with complex numbers, complex-valued neural networks, have been proposed and several studies on their abilities of information processing have been done. In this paper we consider a class of fully connected complex-valued neural networks and discuss existence conditions of energy functions for the networks, analogous to those of real-valued Hopfield type neural networks. Based on the conditions we propose an energy function for the complex-valued neural networks. It is also shown that, similar to the real-valued counterparts, this energy function enables us to analyze qualitative behaviors of the complex-valued neural networks. As an application of the energy function, a synthesizing method of complex-valued associative memories is also discussed.