Soft sensors have been widely used for process control in industrial plants to estimate difficult-to-measure process variables online. A genetic algorithm-based process variables and dynamics selection (GAVDS) method is one method used to select important process variables and optimal time-delays of each variable simultaneously. However, the GAVDS method cannot handle a nonlinear relationship between X and an objective variable y because linear regression is used as a modeling technique. We therefore proposed a region selection method based on GAVDSand support vector regression (SVR), which is a nonlinear regression method. The proposed method is named GAVDS-SVR. We applied GAVDS-SVR to simulation data having high correlation between close pairs of X-variables and a nonlinear relationship between X and y. The GAVDS-SVR method could select regions of X-variables appropriately by considering the nonlinearity and could construct predictive models with high accuracy. Through soft-sensor analysis of industrial polymer process data, we confirmed that predictive, easy-to-interpret, and appropriate models were constructed using the proposed method.

View full abstract

This paper describes aftertreatment temperature control of automotive diesel engine based on reference governor (RG). The RG algorithm computes a modified reference of a catalyst temperature so that the predicted output of the closed-loop system on a finite prediction horizon does not exceed the constraint of its maximal temperature. We apply a bisectional search algorithm to the RG in which the search region is extended in order to take account of model discrepancy. We show the effectiveness of the present method through an experiment with a production vehicle.

View full abstract

This paper deals with a distributed control based on gradient methods for load frequency of power networks including distributed generations, batteries, and renewable energies. The control objective is to minimize the cost function of load frequency control problem of power networks and to operate power systems optimally by means of distributed control based on gradient methods with Armijo and Wolfe type line search. Finally, simulation results of power networks including distributed generations show the effectiveness of the load frequency control.

View full abstract

In this paper, we focus on smoothing effects for photovoltaic generation and investigate its maximization problem based on game theory. We first confirm the smoothing effects through data analysis of global solar radiation profiles. We next formulate an optimal energy source selection problem and then reduce the problem to so-called resource allocation games. After pointing out that the game is reduced to so-called potential games, we present a real-time implementation method of a payoff-based learning algorithm leading players to the optimal action without prior knowledge on utility functions. Finally, we demonstrate its effectiveness through simulation.

View full abstract

We develop a control system to reduce peak power demand by using the internal batteries of office notebook PCs. This system plans a recharging schedule that takes into consideration a demand forecast and user usage, it can handle a wide range of power demand fluctuation without loss of usability. In this paper, we present a synopsis of the system and a local search algorithm to solve the optimal control problem to plan an optimal recharging schedule.

View full abstract

This paper describes synthesis of continuous-time robust deadbeat control for automatic guided vehicle (AGV) in semiconductor manufacturing. Since the dynamics of AGV depends on the loading weight of semiconductor wafer, it is necessary to consider control synthesis which focuses on the change of the loading weight for high-accuracy positioning. In addition to this, the big overshoot and the saturating servo motor current of AGV should be avoided. Motivated by above, first, this paper provides the modeling of AGV. Second, this paper describes robust deadbeat control designs. We discuss the design of the cascade control system which is composed of the servo controller and the deadbeat one. Finally, this paper examines the effectiveness of the proposed systems through numerical examples.

View full abstract

We propose a new threshold determination method for power leveling. This method determines the next term threshold based on the residual quantity transition of energy storage coming from power leveling control through the current period. Additionally, this method can make system implementation easy because it does not need complicating demand forecast and system modeling. We have evaluated power leveling control applied our method by simulations and experiments. And the results showed the efficiency of our method.

View full abstract

This paper presents stability analysis of bilateral control for single-link flexible master-slave arms with constant time delay. The considered system consists of a rigid master-arm and a single-link flexible slave-arm which are connected over the network communication with constant time delay. In this paper, the PDS (proportional, derivative and strain feedback) based bilateral control law for our flexible master-slave arm is employed. The stability of the proposed system is shown via the Lyapunov stability theorem. The performance of our bilateral control is demonstrated by several numerical results.

View full abstract

This paper presents the result of the high performance prop-hanging flight control using quaternions. Prop-hanging flight is known as a way of flight for the fixed-wing aircraft, and it can achieve hovering without special equipments to change the direction of the thrust. We had already succeeded in stabilizing the aircraft during prop-hanging flight using LQI controllers based on the mathematical model. However, a high degree of freedom flight control could not be performed because Euler angles which may cause gimbal lock were used to represent the attitude of the aircraft. In this paper, we utilize quaternions to achieve a high degree of freedom flight control during prop-hanging flight. The effectiveness of the proposed method is evaluated by experiments.

View full abstract

In order to help students understand the basic control theory, the authors have proposed a motivational lecture, where students get personal experience in advance of theoretical lectures. To realize the motivational lecture, the authors have designed a portable and inexpensive control experimental system. However, the system has some points to be improved in terms of costs, size and functions. Hence, the authors revise the experimental system to be more portable and less expensive. Moreover, the system is expanded so that education of other fundamental concepts is possible. This paper shows the design details of the experimental system.

View full abstract

This paper investigates a design of decentralized H₂ static output feedback control for multi-channel linear time-invariant stochastic systems governed by Itô's differential equation. Necessary and sufficient conditions based on cross-coupled stochastic algebraic Riccati equations (CSAREs) for the existence of the strategy set are derived. In order to prove the usefulness of the proposed methodology, the extension to weakly-coupled large-scale stochastic systems is also considered. It is shown that the parameter-independent strategy set can be designed by solving the reduced-order stochastic algebraic Riccati equations (SAREs). Furthermore, as a novel contribution, the degradation of H₂ norm for the closed-loop stochastic systems by means of the parameter independent strategy set is discussed. A numerical example is given to demonstrate the useful feature obtained.

View full abstract

This paper deals with a leader-following formation navigation (LFFN), which is one of the most important issues in formation control. In this formation, one leader is controlled by a human operator, while multiple followers pass along the leader's trajectory in the same motion to the leader. The followers move along the leader's path behind the leader after some intervals. In this navigation method, the distances between the robots are not considered, thus some of the followers are possibly too late to maintain the formation. This paper proposes a new LFFN method with the maintenance of the formation-distances. The proposed method minimizes the sum of the evaluation functions concerning the velocity and distance errors. The effectiveness of the presented method is demonstrated by an experiment with three mobile robots.

View full abstract

This paper presents a super-resolution process control system for maintaining good super-resolution performance. Our idea for the maintenance is that the already existing head control system is developed by adding head's rotation velocity constraint and reference governor to guarantee acquisition of the specified number of images used for the super-resolution process. An effectiveness of the development is confirmed via experimental verification, compared to our previous work.

View full abstract

This paper considers a feedback engine speed control method for ships in head regular waves. We evaluate the fuel efficiency of our controller against the case of constant throttle opening through numerical simulation.

View full abstract

Bridge pier joints occasionally become unstable due to deterioration and after the great earthquake disaster. The purpose of this paper is to report on a measurement of vibration characteristics for the bridge using piezoelectric cable. The cable winds up the piezoelectric film in the surrounding electric wire. It also generate high voltage for amplifiers when the cyclic loading is applied, which does not require instruments because it has a piezoelectric. The output voltage was generated in proportion to the fluctuation of the vibration.

View full abstract