This paper introduces an elevator design support system which employs case-based reasoning techniques. The system analyzes elevator specifications, generates indices, retrieves similar cases from a case base, and shows the retrieved cases to elevator designers. The biggest issue in applying case-based reasoning techniques is retrieval of appropriate cases. For the appropriate retrieval, indices of the cases and the definitions of similarities among the cases are correctly chosen. In elevator specifications with many items, it is extremely difficult to make the whole structure of the specifications clear for choosing indices and defining the similarities among the specifications. In order to overcome the difficulty, a new indexing method which uses heuristics of the designers is developed. Also, the system provides the designers with flexible retrieval ways, so that the designers can refer the cases from various points of view.
A method is presented for detecting self-similar patterns. A version of conditional distribution is introduced to self-similar patterns for representing computable morphological feature. The morphological feature associated with accurately observed attractor is proved to yield a subset that is invariant with respect to self-similarity mapping. The invariant feature is demonstrated to activate an imaging process for generating proximately located attractor. An algorithm was developed for computing invariant feature and verified through simulation study.
Recently, traffic accidents have been increasing on the Hanshin expressway and the number of the accidents already exceeds 6, 500 in a year. Once an accident occurs on the expressway, it is likely to lead to the temporary closing of lanes and a potentially serious disaster by the following vehicles. If accidental events, such as accidents or stopped vehicles, could be detected promptly, the alarm of the events to the following vehicles and to the traffic control center would become faster than it was. Feasibility studies and their experiments have been conducted, whether the events can be detected by processing images taken from video cameras. The prototype system was installed at a curved area on the expressway where the events frequently occur. The study of the detection algolithms and the verification of the effects after the prototype system installation will be reported in this paper.
Recently, the author proposed a new learning algorithm derived algebraically which is composed of the error back propagation to give a fictitious teacher to each hidden layer and the update rule of each connection. The details of the algorithm with some comments are demonstrated in this paper using the Exclusive-OR problem as an example. Some update rules of each connection are also proposed here. Among them, the update rule using an orthonormalization method with some scheme is most effective. Simulation results show thst 5 steps convergence can be obtained in many cases of different initial values. The exceptional cases which do not show the convergence are also made convergent by some improvement in the nonlinear characteristic of the unit.
In this paper, we propose a subspace state-space system identification method considering noise attenuation. In this method, we achieve noise attenuation based on the uncorrelation between input signals and noises, where a large amount of data can be handled with the prescribed size matrices. An illustrative numerical example is given to show the effectiveness of the proposed method.