This paper deals with a prediction method of air pollution levels by using the space-time series model which considers both time lag and spatial lag to introduce the concept of spatial spread. As it is considered that the prediction accuracy is influenced by the wind direction, We propose a new method to take the information of wind direction in the model. From the simulation results, it is shown that the prediction accuracy based on the space-time series model is superior to that based on the ordinary time series model which considers time lag only.
An application of a model-based control system to the V-bending process is studied. Since the accuracy of the products is affected by the scatter of the material characteristics, the accuracy should be improved if the punch stroke is adjusted properly according to the identified material properties for each workpiece during the process. A simulation of V-bending process is first performed to investigate the deformation behaviour of the material in bending and the effect of the scatter of the material properties on the deformation. The model used in the simulation is then utilized in the control system to calculate the punch stroke. A bending apparatus is built and the identification of the material properties is performed on-line to cope with the scatter. It is experimentally verified that the products with higher accuracy are obtained by the newly developed process than by the process without the model-based control.
This paper discusses image registration on the basis of the node structures which are represented by the arcs linked to them and their shapes of the image graph and its application to stereo visions (motion stereo and binocular stereo vision). It is necessary but difficult to find matching between feature points on a pair of stereo images when a vision system extracts three dimensional positions and shapes of objects. This paper proposes a matching manner using an image database in the case when the occlusion occurs on the pair of stereo images. Since our approach is based on an algebraic expression of edge line image using unit vectors and their concatenation operators, it is insensitive to scaling, rotation and gray level modification between the pair of stereo images. The matching process makes use of a degree of structural similarity between nodes of the image graph. The experimental results confirm that our approach is effective for the image registration between the pair of images including the occlusions.
In order to control a large-scale distributed process, it is important to predict output variation. A rule-based prediction model and a statistical one have been developed. Applying the fuzzy inference theory, we have connected these models and developed a total prediction model. In addition to these approaches, we have developed a heat level evaluation system using fuzzy inference. This system is used for evaluation and maintenance of a heat level prediction system. Using the predicted result, the action guide system to control the blast furnace heat level has been developed. By using this guidance system, the heat level can be controlled more accurately than before.
This paper describes the CAI system for indefinite integral. The system has five fundamental functions for intelligent tutoring system as follows : 1. Interactive teaching, 2. Problem solving, 3. Error detection, diagnosis, and remedial action, 4. Explanation for WHY and WHAT questions, and 5. User-friendly interface by mouse and window system. And students can use the system on 16-bit personal computer. The system consists of the five parts such as “interface”, “input analysis”, “student model”, “pedagogical strategy”, and “knowledge base”. In the interface, there are four modes : input, request of hints, questions and termination modes. The student can select each mode according to his understanding. In the input analysis, a student who selected the input mode in the interface part inputs the solving procedure, e. g. “I solve the problem in an indefinite integral”, or “I use the distribution law”. Then he inputs a solution of the problem. The validity of the student's solving procedure is evaluated by comparison between his procedure and that created by the system. The validity of the student's solution is evaluated by comparison between his and solution according to the knowledge base of the system. Then the system diagnoses student's understanding based on the both validities. The student model consists of seven lists such as  list of student's understanding about solving procedure,  list of student's understanding about solution,  list of solving skills,  bug rules,  number of careless mistakes,  number of hints required by student and  documents of input of student and output of the system. In the pedagogical strategy, the system gives the student some hints and easier problems to get the initiative from him. The knowledge base is written in inference rules of the prolog style and is constructed hierarchically. Thirty-two of high school students solved the ten problems before/after use of the system. The results of the tests before/after showed 82.5 % of unacquired knowledge was aquired by use of the system.