Security of Cyber-Physical Systems (CPS) against cyber attacks is one of the most critical issues that should be investigated. In this paper, we deal with the vulnerabilities due to novel integrity cyber attacks named perfect stealthy attacks and the countermeasure against them in CPS which are modeled as a stochastic Linear Time-Invariant (LTI) system equipped with a Kalman filter, an LQG controller, and a χ² failure detector. The attacks are designed by a sophisticated attacker so that the measurement residual of the compromised system coincides with the healthy one, and can ruin the system definitely. First, from an attacker's standpoint, we show an existence condition and a design procedure of the attacks using invariant subspace property. Then, on the other hand from a system supervisor's standpoint, the countermeasure against the malicious attacks using an aperiodic transmission scheme based on the Nash bargaining game is discussed. Finally, the dangerousness of the attacks and the effectiveness of the proposed detection method are verified through numerical simulations.

This paper proposes a motion planning method to manipulate an grasped object by a two-fingered robot hand based on the nonholonomy of the rolling contacts. A cubic grasped object is regulated to a target configuration through regulating the contact coordinates between the fingers and object using the holonomy generated by closed paths on spherical finger-tips. The parameters of the closed paths are determined by solving an optimization problem with the constraints which consist of the limitations on the finger-tip/object surfaces and the finger joint configurations. The proposed method is verified by a numerical simulation, where the planned motion is impossible without using the nonholonomy.

In order to alert in advance car drivers to the danger of traffic accidents caused by the slipping of tires on frozen road surfaces, this report presents an identification method that can discriminate the frozen surface and wet one in high accuracy. Being different from the conventional methods of using information about wet condition and low temperature on the road, we propose another discrimination method that uses only quasi-millimeter waves. The report focuses on the phenomenological fact that the reflection of the waves from ice surface is greatly different from water one. For confirming such theoretical prediction, we carry out experiments in the laboratory using a portable and commercially available transmit/receive module with a 24GHz frequency band. The reflection coefficients are obtained by measuring standing wave ratios. Experimental data show that the reflection coefficients from the water film are stably greater than those from ice film on average. Thus obtained results demonstrate that the present method provides the feasibility of discriminating the water and ice surfaces on the road in actual environmental situation.

This paper proposes an airborne ultrasound transducer driven with electrostatic force, which potentially enhances applications of airborne ultrasound. The proposed transducer can be fabricated by laminating flexible materials and can be formed as a sheet-like phased array device. As the first step, we fabricated a prototype of a single transducer. In experiments at 44kHz, we obtained 55% electro-acoustic energy conversion efficiency and 8.7Pa sound pressure at 20cm from the circular sound source of 6.4mm in diameter. These basic properties suggest that the proposed method enables a thin and light weight airborne ultrasound phased array that is usable in nonlinear acoustics applications.

When a group of people orders dishes at a restaurant, in many cases it is desirable to realize “simultaneous offering of dishes” - to carry all dishes as simultaneously as possible. However, it is difficult to manage the cooking schedule so that the dishes ordered by the same group are completed at the same time in a situation where many other dishes are ordered by other groups. In this paper, we develop a scheduling system to generate a cooking schedule to realize simultaneous offering of dishes in accordance to the current situation on the orders from all groups. We also propose a dynamic scheduling method to generate a schedule in dynamic environments where orders come in one after another.

Nowadays shipping industry has problems such as reduction of mariners and aging of mariners. In order to solve the problems, One Person Bridge Operation was developed as navigation supporting system. However, One Person Bridge Operation remained one man error and technological problems. Therefore, in this paper, remote maneuvering between land and ship was proposed by using satellite communication. In this remote maneuvering system, navigation officer works on ship as usual and helmsman supports ship from land. In this study, remote maneuvering system was developed by using satellite communication. For evaluating the remote maneuvering system, ship experiment of remote maneuvering was carried out by navigation officer and helmsman. From the result, navigation officer and helmsman could maneuver a ship accurately between land and ship.

When applying a reinforcement learning method, the estimation accuracy of the state transition probabilities affects the performance of the policy obtained from the estimated plant. Therefore, we find a sampling condition guaranteeing that the optimal policy from the estimated plant is also optimal for the real plant with the desired degree of reliability, and a sampling methods based on it is proposed. Not by the sampling for the reliability in which the policy is optimal for the real plant, but by the sampling for the policy to be effective irrespective of estimation errors, we can further reduce the number of samples. We show the problem setting for finding the policy which is guaranteed to be effective for estimation errors from the real transition probabilities with the desired degree of reliability, and we propose a sampling method as a solution of this problem. The effectiveness of the proposed method is verified by numerical simulations.

This paper proposed a walking assistance method by using trunk rotation based on the spinal engine theory. According to the spinal engine theory, our walking is realized by cooperation between upper body and legs through the energy exchange. Therefore, we developed an assistance device for trunk rotation and its control method. We adopted the periodic input control as the control method and modified it to be suitable for human assistance. According to experimental results, the proposed method realized walking assist by supporting trunk rotation movement. Moreover, it also realized training effect as the walking resist, by controlling a parameter of the controller. In this article, their effects were confirmed by theoretical analysis and experiments.

In this paper, we proposal a method to estimate comfortable clothes depending on thermal environment. We adopt a thermal environment model in order to estimate thermal comfort of a human. This model consists of a weather information block, a solar radiation model, a dress database, and a human thermoregulation model, and it is possible to calculate Standard New Effective Temperature (SET*) by adopting the thermal environment model. The thermal environment model evaluate the thermal comfort based on the calculated SET*, and estimate comfortable clothes.