Many traditional cultural assets in Kumamoto City were damaged by the Kumamoto Earthquakes in 2016. The Kumamoto Castle disastrously damaged by the earthquakes should be completely repaired because it is one of the most important cultural assets in Japan. The Kumamoto Castle recently celebrated its 400th anniversary. In this paper, we propose database design method of broken stones in the Kumamoto Castle for completely repairing the wall of stones which is called “ISHIGAKI”. It is required to find accurate position of each stone before the Kumamoto Earthquakes. However it is difficult to matching each stones because of the shooting environment problem of the photographs of each stone. It is required to construct ISHIGAKI database for matching collect position of stones. We focus on location after broken ISHIGAKI and the surface area of each stone for increasing the ability to find the collect location. Moreover, we propose the entire ISHIGAKI matching algorithm based on not only image matching but also surface area information and location information of stones.

A multi-rate state observer with a periodically time-varying gain is proposed in this paper. The proposed observer can be regarded as a periodically time-varying system. By using the cycling technique, the periodically time-varying system can be dealt with as a time-invariant system. This paper presents a method to design a multi-rate state observer using the cycling technique. Then, analysis about the l₂-induced norm of the estimation error by the proposed observer is reduced to an LMI optimization problem. Then, the design problem of the time-varying gain is also described as an LMI optimization problem. The effectiveness of the multi-rate state observer is shown by numerical examples.

Model based control system design has been developed in many fields based on the dynamics of the accurate plant model. The desired control performance of the transient response can not be achieved if the dynamics gap between the plant and its mathematical model is not small. We have proposed a specified compensator to minimize the gap between an actual plant and its nominal model. The proposed compensator is composed of its plant model and an error feedback component. In this paper, we propose a method to analyze error feedback compensation component based on the analysis method using LMIs. Polytopic-type uncertain continuous LTI systems are dealt with in this paper. The robust performance by the model error compensator is analyzed using H₂ and H_∞ indexes. The effectiveness of our analysis method is shown by numerical examples.

In this paper, we propose a pose synchronization law on SE(3) under disturbances. First, we present a control law which uses partial information of disturbance models. Then, we show that rigid body motion has incremental passivity-like property and prove that the proposed control law achieves disturbance rejection with pose synchronization on SE(3). We also show that our results are reduced to disturbance rejection with attitude synchronization on SO(3) with only relative information. Finally, we demonstrate the validity of the control law by simulations. Additionally, we build a planar (2-D) testbed and show its efficiency.

The bounded real lemma for deterministic linear time-invariant systems is known as one of the most important tools in robust stability analysis. A counterpart for systems with stochastic dynamics is considered to be also important, and A.E. Bouhtouri, et al. (1998) have derived it for discrete-time systems with state-dependent noise. However, the class of the systems is restrictive, and has room for generalization. The purpose of this paper is to show the sufficiency for a new bounded real lemma for more general systems with stochastic dynamics determined by an i.i.d. process. The validity of the lemma is demonstrated with numerical examples.

This paper dealt with a design method of MCV observer using robust invariant set analysis method. In our previous research, a design method of state observer for output signal with outliers was proposed. In some control systems, adverse effect by outliers is one of the most important topics for solving. To overcome the adverse effect by outliers, we proposed a method to disappear the effect of outliers using median and candidates of estimated state. In this paper, we give an evaluation signal, and analyze the maximum absolute value of the evaluation signal using a robust invariant set. We propose the design method of observer gains to minify the maximum absolute value of the evaluation signal based on LMIs with some variables. It is easy to design the observer gains by our method. The effectiveness of the proposed design method of observer gains is evaluated by numerical simulation.

In this paper, we propose a high-resolution ranging method for sensor nodes in wireless sensor networks based on RF TOF measurement. Ranging methods using TOF have a high-accuracy because TOF of RF signal is hardly affected by surrounding conditions. However, it is difficult to measure with enough resolution due to its very fast speed. In our laboratory, the principle of vernier caliper, which is called vernier effect, has been used to solve the problem. By using the principle, finer resolution than the periods of measurement clocks can be obtained. As problems of the method, however, it has been reported that the measurement takes too long time and requires additional circuits which lead to higher power consumption and cost. Then, we propose a novel method in this paper for obtaining TOF with just simple calculations and no extra hardware. In this method, the correlation property of M-sequence code is used to achieve accurate and easy measurement. By transmitting and receiving M-sequence code with different clocks and conducting a devised calculation of correlation value, TOF is accurately measured with high-resolution. Also, some experiments were conducted in our research. In the first experiment, we measured the delay time on delay lines. From the result, we found the condition that the measurement results may get worse and the way to achieve enough resolution even in the condition. In the second experiment, the proposed method was applied for laser ranging and it was confirmed that distance can be measured with 3m resolution.