For motor controlled Euler-Bernoulli beam without a tip body, it has been shown that strain feedback can exponentially stabilize the whole system. On the other hand, it is observed in the existing literature that the exponential stability may no longer hold under boundary velocity feedbacks if the beam is fitted with a tip body. So care must be taken in dealing with equations of beam with tip body. It is thus the purpose of this paper to investigate the exponential stability of the strain feedback controlled dynamic equation of an Euler-Bernoulli beam with a tip body. The energy multiplier method is employed in proving the main theorem in which a key point is to choose an appropriate multiplier.

View full abstract

Recently, much research on application of neural network (NN) using genetic algorithm (GA) has been reported. In this paper, we apply GA to a neuro-pattern recognition system for paper currency with masks. We regard the position of the masked part as a gene on the input image. We operate crossover, mutation, and selection to some genes. By repeating a series of these operations, we can get effective masks for recognition of paper currency. We compare the ability of NN using the optimized masks by the GA with the one of NN using the random masks. Then we show that the GA is effective to mask optimization for the method of neuro-pattern recognition. Furthermore, we refer to a high-speed neuro-recognition board to realize the neuro-pattern recognition for paper currency in the commercial products.

View full abstract

Error back propagation is one of the most popular ideas used in learning algorithms for multilayer neural networks. Since the pioneering work by Rumelhart, Hinton and Williams, error back propagation has been regarded as a gradient descent method or its close approximation to minimize the sum-squared error function. In this paper, we point out that “on-line” back propagation is better interpreted as a successive projection method for solving a system of nonlinear ineqnalities. In particular we proposed a new learning algorithm based on the successive projection method, in which the stepsize is determined quantitatively based on the magnitude of error observed for each input pattern. Some simulation results on XOR and parity check problems indicate that the propsed algorithm is more effective and robust than the standard on-line back propagation algorithm.

View full abstract

In this paper, a method to compute almost exact gain margin of uncertain control systems using Polygon Interval Arithmetic (PIA) is proposed. The main problem here is to determine 0-exclusion property of the value set of the characteristic polynomial. By using PIA, we can compute the convex hull of the value set very fast. When we need a better estimate of the value set, we need to split a region of uncertain parameters : however, the method to decide which region should be split was not given, and it may require very large computing time to get a good estimate. In this paper, we confine ourselves to the case when an expression of characteristic polynomial which corresponds to a totally decomposable tree structure decomposition is given, and we propose a method to decide which region should be split. By adopting this method, we Can compute a good estimate in a short time.

View full abstract

In this paper, we study supervisory control of discrete event systems under partial observations. Most of the research on supervisory control considers a supervisor which takes a control action according to only event sequences. This paper proposes a supervisor which selects a control pattern based upon partial observations of events and states. We first derive necessary and sufficient conditions under which there exists a supervisor for a given language, namely, a control specification. It is shown that our supervisor is more powerful than a supervisor which uses only event sequences. Next, we study the case that there does not exist a supervisor for a given language. We present a sublanguage of a given language for which a supervisor exists.

View full abstract