研究発表

This paper deals with a dynamic pricing via the H-infinity control considering uncertainties in market participants' behavior. The dynamic pricing is a decision procedure of an electricity price based on power demand and supply information from selfish power consumers and generators. However, since market participants have uncertainties in their behavior, the price decision procedure is required to have robustness against these uncertainties for stable operation of power systems. Then, in this paper, we propose a novel price decision procedure via the H-infinity control and also show the effectiveness of our proposed price decision method in simulation results.

This paper discusses faults diagnosis method of multiple cyber attacks in networked electrical power systems. We deal with a power network of a centralized system, and then, some parameters are estimated by Kalman filter. Using a sensor network system, each of sensor nodes can exchange information by wireless communication. A fault diagnosis method is proposed by using both fault detection diagnosis matrix and fault distinction matrix. The former is composed of estimated values and observed values. On the other hand, the latter is composed of observed values and calculated values via the senor network. Finally, the effectiveness of the proposed approach is validated by a simulation experiment.

This paper proposes a novel direct load control method to stabilize the system frequency. The proposed method utilizes the consumers' controllable loads for load frequency control. By solving an optimization problem to maximize the social welfare function, the input required for the frequency control is allocated among the controllable loads and the load frequency control generator in an economically rational manner. To introduce the direct control program for smooth operation of the power system, preventing consumers from cheating and guaranteeing their benefits are required. Thus, strategy-proofness and individual rationality based on the Groves and Clarke mechanism were used to design the electricity charge. Each consumer bids on the utility functions truthfully to the independent system operator and gains from participating in the proposed direct load control program. The effectiveness of the proposed method was demonstrated through numerical simulations.

This paper addresses wind power prediction which is known to be a key technology in EMS(Energy Management Systems). In this paper, 24 hours ahead power prediction method using a filtering theory is proposed for wind power generation. The prediction method is a simple algorithm, the procedure of prediction consists of two steps, the data processing and the calculation of predicted values. In the calculation of predicted value, the unknown parameters are estimated via constrained kalman filter. Moreover, in the estimation of the unknown parameters, the convergence of them are also guaranteed. Finally, the advantages of the proposed method over the conventional method are shown through actual prediction evaluations.

This paper addresses wind power prediction which is known to be a key technology in EMS(Energy Management Systems). In recent years, an introductory expansion of renewable energy is expected and the prediction of wind power generation is needed for taking in wind power generation. The goal of this work is to predict the amount of generation in the next day from the past actual data and the weather forecast data of wind. In this paper, 24 hours ahead power prediction method using a filtering theory is proposed for wind power generation. The prediction method is a simple algorithm, the procedure of prediction consists of two steps, the data processing and the calculation of predicted values. In the data processing, in order to get the correlative data from the database, we employ JIT(Just-In-Time) Modeling. In the calculation of predicted value, we provide the regression model for wind speed and wind power, and the unknown parameters are estimated via constrained kalman filter. Finally, the advantages of the proposed method over the conventional method are shown through simulations.

本講演では，エネルギーネットワークシステムの分散的管理・制御問題に対して，分散協調制御理論からのアプローチとして，講演者のこれまでの一連の研究成果を紹介する．具体的にはスマートグリッドの分散制御，電力市場における分散的最適価格設定法と需給制御，スマートグリッドの最適エネルギーマネジメントに関して議論するとともに，その基盤となる協調制御，分散制御に関する過去の研究についても言及する．最後に今後の研究の展望と将来の方向性を述べる．

In this study, we consider cooperative control issues for a multi-unmanned aerial vehicle (UAV) system. Specifically, we present a cooperative formation control strategy for a multi-UAV system with unidirectional network links. Our strategy is to apply a consensus-based algorithm and leaderfollower structure to the UAVs so that they can cooperatively fly in formation. The leader provides each UAV with commands to generate a geometric configuration of the formation. Convergence is guaranteed when the cooperative formation control algorithm is applied to the UAVs. Collisions among UAVs can occur when they are flying with the cooperative control UAVs. Our strategy for collision avoidance is to apply an artificial potential approach to the UAVs. Experiments are performed on multiple commercial small UAVs to validate the proposed formation control algorithm with collision-avoidance capability.

This paper deals with a robust formation control methodology for the Multi-UAV system with uncertain exogenous disturbance. The fourth-order UAV model with disturbance and the disturbance estimate system is introduced. In order to achieve the rigid formation fight, graph theory and consensus algorithm are utilized and the state of the UAVs converges to desired value by using Leader-Follower structure. The proposed control law includes the Robust Integral of the Sign of error (RISE) for estimation of the disturbance and also involves collision avoidance property based on potential method. The stability and convergent properties are proven based on Lyapunov theory. Finally, several simulation results show the effectiveness of the proposed method compared with a conventional sliding mode control.

This paper deals with distributed and cooperative control algorithm to make each agents to form a formation for a second-order system. However, when forming a formation with a multi-agent system, there is a possibility that the agents collide with each other or collide with obstacles other than the agent. Also, if the agents avoid other agents and obstacles, the distance between agents may be temporarily too far apart. In order to solve these problems, this paper proposes a control algorithm which consists of Leader-Follower structure, consensus algorithm and artificial force based on artificial potential field. For the control algorithm to make agents achieve formation in a decentralized cooperative way, we derive the condition concerning the control gain to ensure stability. In addition, by comparing the amount of energy, we design conditions that ensure collision avoidance between agents, collision avoidance with obstacles, and distance between agents, and also design potential functions that satisfy the conditions. The validity of the control algorithm proposed in this paper is verified by numerical simulations.

This paper deals with the scenario of two multi- UAV systems (so-called teams), each flying from a starting point to a target point. The teams meet on their ways and have to avoid collisions in the horizontal plane. Both teams work independent of each other but use the same formationcontrol and obstacle-collision-avoidance algorithm. The leaders measure obstacles' positions, use a Kalman Filter to estimate their movement and predict their future positions. With the information about the other agents' positions each agent calculates its reference trajectory. The leader also computes a trajectory to guide the team towards the target point. Distributed MPC is used for each agent to track its reference trajectory. If there is a risk of colliding with an obstacle, the MPC cost function is extended by a penalty term to ensure obstacle collision avoidance. The stability of formation control is proved in this paper. The capability of two teams avoiding collisions is shown in a simulation.

This paper proposes a RISE-based cooperative control of uncertain multi-agent system with exogenous disturbances. First, we introduce the second order model with disturbance for agent. The network topology among agents is undirected and connected, and information of reference is allowed to be available to at least one agent. Second, we propose the control law consists of graph theory, consensus algorithm and RISE. Graph theory and consensus algorithm are utilized to converge the states of agents to reference value, and RISE is utilized to suppress the disturbance. Then, we show the proposed control law guarantees the agents achieve synchronization, and derive the condition to achieve the control objective. Finally, simulation results show the effectiveness of the proposed control law.

Ensuring cyber security of smart grid based on supervisory control and data acquisition using information and communications technology has been becoming a major challenge. Especially, a replay attack is a dangerous cyber attack affecting integrity and authorization. The goal of this paper is to detect replay attack that is one of cyber attacks on the sensors of a control system. We propose a detecting method adding intentional noise to not only sensors as code signal but also input. Replay attack can immediately be reflected by using fault diagnosis matrices that are composed of the estimator and observed values even if code signal is decrypted. Finally, we show simulation results to analyze effectiveness of the proposed method.

This paper deals with a network of interconnected micro grids. The transient dynamics is modelled as an averaging process involving dynamic agents in a network. An analysis of the convergence of the consensus dynamics is provided using a network model based approach and by exploiting the properties of the corresponding graph-Laplacian matrix. Furthermore an investigation of the transient dynamics is carried out for different damping and inertial parameters and under different timevarying topologies. Finally a simulation is performed based on a model calibrated on an existing network in the UK under parameter uncertainties.

This paper presents a novel optimal power demandmanagementmethodforpowerconsumerswithanaggregator considering state and control constraints. The aggregator is an organization who manages some consumers to achieve demand response efficiently. This paper proposes a power demand adjustment method for the aggregator to allocate power reduction among consumers in a distributed manner. Specifically, this method simultaneously derives the optimal power reduction of each consumer using a gradient method and determines the control inputs by solving a quadratic programming problem with state and input constraint conditions. This paper also shows the convergence of the proposed algorithm, and finally, numerical simulation results illustrate the effectiveness of the proposed method.

本研究では磁気浮上システムに対して漏れインダクタンスを考慮したモデルを考え，モデルを電気系サブシステムと機械系サブシステムに分割し, その受動性を示した．またそれぞれのサブシステムに対してのコントローラを設計した．制御実験ではステップ応答, さらに浮上体の質量を変更したときの応答を観測し，提案法では従来法に比べ過渡応答時の振動を抑えることができた．

本稿では，GIMC 構造が高性能とロバスト性を両立できることを実験的に示した．GIMC 構造は高性能なコントローラと高ロバストなコントローラの二つを内部信号によって切り替えることができる．実際にGIMC 構造を磁気浮上システムに応用し，H∞制御と比較検討することによって，高性能ロバストコントローラが実現できることを実験的に示した．

2010 IEEE MSC （システムと制御に関する合同会議）において，本会議併設行事として，広く，公開市民講座を開催した．具体的にはH∞フィルタに基づく電力需要予測とマイクログリッドの分散予測制御について述べた．需要予測のためにH∞フィルタを用いて電力需要特性のパラメータ推定を行い，これによって, モデル化誤差に対し，より正確な推定が可能となったことを平易に解説した．さらに実際の運用を想定し平日と休日の需要曲線の特徴を考慮した電力需要予測手法について詳しく説明した．