In this paper, first of all, the group walking behavior by four pedestrians are observed. In the observation, not only the motion data but also the decision making of each pedestrian are collected by using special device. Then, three behavioral indicators: deceleration, detour amount, and decision entropy, are defined and calculated. It has been found that these three indicators successfully quantify the ‘smoothness’ of the group walking behavior. Finally, the principal component analysis (PCA) is applied to the three dimensional indicator data. As the result, the meaning of three principal components are clearly explained. The discussion based on the PCA will be a basis for the further analysis and classification of the group walking behavior.

In the operation of brushless DC motors, a method to observe angular velocity from hall sensor signals is widely used. However, the observation accuracy deteriorates due to misalignment of hall sensors and magnetic poles on the rotor. To solve this problem, Kalman filtering based on the observed signals is effective to estimate angular velocity. In order to improve the estimation performance, the proposed method adopts a parameter setting for the variance of the measurement noise in Kalman filter algorithm, which is set to be depending on the estimated angular velocity. This paper shows that the variance of the measurement noise is obtained through a simple preliminary experiment and the effectiveness of the proposed method is verified by experimental results.

This paper proposes a state estimation scheme over lossy bidirectional communication channels by means of a UIO (unknown input observer) and gain switching. Under some conditions on loss patterns and on system matrices, the proposed scheme guarantees that a weighted square norm of the estimation error either decreases monotonically, or stays below the latest-received value, if observation is received or lost, respectively. The switching gains are designed by solving LMIs (linear matrix inequalities) simultaneously.

In this paper, we propose a new algorithm that adjusts the parameters of the distance function during dimensional compression according to the degree of influence of each component on the object variable. In this method, lower dimension space is divided into subspaces. And linear multiple regression to a predetermined objective variable is executed for each subspace. From result of this regression, weight parameters of the distance function are updated. By applying this method to the self-organizing maps using the manufacturing conditions of the hot rolling process of aluminum as input, it was confirmed that appropriate and clear dimensional compression results could be obtained. Furthermore, quantitative evaluation confirmed that the arrangement in the distance space after dimensional compression are separated for data with large differences in the object variables. This method is an algorithm that can be applied to kernel PCA and t-SNE that do not use distance functions, and is expected to be applicable to other general high-dimensional data.