Takehito Azuma

Development and Experiment of Networked Control Systems with Congestion Control

The purpose of this paper is to develop an autonomous vehicle system including communication networks and to demonstrate congestion controllers for networks based on state predictive control via the developed system. In the net- worked autonomous vehicle, the vehicle has one camera as sensor. Images from camera are sent to a computer via computer networks and are processed to recognize circumstances around the vehicle. Using the processed camera image, the vehicle is controlled over computer networks. By considering that this system uses computer networks, congestion problems are important because congestion causes instabil- ity of the whole system. For these problems, congestion controllers are designed based on the systems & control theory. In this paper, new congestion controllers are designed by using state predictive control to assure stability of computer net- works. Effectiveness of congestion controllers is verified in experiments using the developed system.

A New Approach to Estimation of Protein Networks for Cell Cycle Based on Least-Squares Method

This article considers a method to estimate protein networks for cell cycle based on system identification in control engineering. The approach provides a new approach to estimation problems of protein networks for cell cycle in systems biology. First the considered estimation problems of protein networks are defined. Second our approach based on least-squares method is shown. Finally by applying the proposed approach in a numerical example, a protein network of 6-dimensional cell cycle systems is estimated and this network consists of some known protein network for 6-dimensional cell cycle systems.

Estimation and robustness analysis of protein networks for cell cycle systems

This paper proposes an estimation problem of protein networks for cell cycle and their robustness analysis. Our proposed method is explained to estimate a protein network for cell cycle based on system identification techniques. The method is based on the least-squares method for state space models. Applying the method, a 6-dimensional protein network is estimated for cell cycle to demonstrate the efficacy of our proposed method. The estimated protein network is called as 6-dimensional cell cycle system and the cell cycle system is described as nonlinear differential equations. Moreover robustness analysis and numerical simulations are performed for the cell cycle system described as nonlinear differential equations. From these results, robustness of the cell cycle system is discussed.

An experimental result on system identification over networks using delta-sigma transformation

In this paper, an method to system identification over networks using 1 bit delta-sigma transformation is proposed and the efficacy of the proposed method is verified based on experimental results. Accurate mathematical models are needed to achieve intelligent control with good performances in control engineering. It is easy to obtain those mathematical models if exact input-output data of a controlled object is available by applying system identification techniques. However it is difficult to obtain the exact input-output data over networks because the data is transformed from analog data into digital data. The proposed method provides a method to build mathematical models of controlled objects over networks.

Understanding of robustness for a basic system of cell cycle in yeast

This paper discusses a robustness analysis of cell cycle in yeast, which is one of eukaryotic cell cycles, and focuses on the understanding functions of Cdc25 and Weel proteins by using the Michaelis-Menten method. The robustness of the cell cycle is analyzed based on the sensitivity analysis for a mathematical model. From the first analysis result, it was shown that Cdc2 and Cyclin proteins have main roles for cell cycle in this model but the robustness is not high against perturbation on its parameters. By introducing Cdc25 and Weel proteins to the mathematical model, it was verified by the sensitivity analysis that the modified has higher level of robustnesses than the old model does. Numerical examples are shown to demonstrate the old model and the new model have almost identical cell cycle behaviors leaving robustness as a salient difference. Finally the difference of the Michaelis-Menten method and the Mass equation method are well understood.