• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.5
• /
• pp.401-406
• /
• 2012

This paper considers multi-agent systems with interval time-varying delays and switching interconnection topology. By construction of a suitable Lyapunov-Krasovskii's functional, new delay-dependent consensus control conditions for the systems are established in terms of LMIs (Linear Matrix Inequalities) which can be easily solved by various effective optimization algorithms. One numerical example is given to illustrate the effectiveness of the proposed methods.

• International Journal of Control, Automation, and Systems
• /
• v.1 no.3
• /
• pp.358-367
• /
• 2003

We consider discrete-time factorial Markov Decision Processes (MDPs) in multiple decision-makers environment for infinite horizon average reward criterion with a general joint reward structure but a factorial joint state transition structure. We introduce the "localization" concept that a global MDP is localized for each agent such that each agent needs to consider a local MDP defined only with its own state and action spaces. Based on that, we present a gradient-ascent like iterative distributed algorithm that converges to a local optimal solution of the global MDP. The solution is an autonomous joint policy in that each agent's decision is based on only its local state.cal state.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.59.5-59
• /
• 2001

Multi-agent robot system is the system which executes by cooperating with each robots and controlling several robots. Capability and function of each robot must be considered for cooperation behavior. Furthermore, it is necessary to analyze the given environment and to replace complex task with some simple tasks. Analysis of the given environment and role assignment for the given tasks are composed of discret event. In this paper, the hierarchical controller for multi-agent robot system using the petri-net state diagram is proposed. The proposed modeling method is implemented for soccer robot system. The effectiveness of proposed modeling method is shown through experiment.

• Journal of Institute of Control, Robotics and Systems
• /
• v.17 no.9
• /
• pp.857-862
• /
• 2011

This paper studies the output consensus problem for a class of heterogeneous linear multi-agent systems under a fixed directed communication network. The dynamics, as well as its dimension, of each agent can widely differ from the others, but all the agents are assumed to have the same transfer matrix. In addition, only the system outputs are constrained to be delivered through the network. Under these conditions, we show that the output consensus is reached by a group of identical controllers, which is designed to achieve the state consensus for the homogeneous multi-agent system obtained from the minimal realization of the transfer matrix. Finally, an example is given to demonstrate the proposed result.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.3
• /
• pp.222-226
• /
• 2010

In this paper we study the consensus problem of SISO (Single-Input Single-Output) linear multi-agent systems under a fixed weighted undirected network communication. The dynamics of each agent can be of any order, while it is constrained to be neutrally stable. Based on a simple root locus argument, we show that the problem at hand is always solvable if the gain and zeros of the dynamic compensator, of which dimension is the same as the number of plant poles on the imaginary axis, are chosen suitably, finally the result is successfully verified by an example.

• Journal of Korean Association for Spatial Structures
• /
• v.23 no.2
• /
• pp.29-36
• /
• 2023

Reinforcement learning (RL) is widely applied to various engineering fields. Especially, RL has shown successful performance for control problems, such as vehicles, robotics, and active structural control system. However, little research on application of RL to optimal structural design has conducted to date. In this study, the possibility of application of RL to structural design of reinforced concrete (RC) beam was investigated. The example of RC beam structural design problem introduced in previous study was used for comparative study. Deep q-network (DQN) is a famous RL algorithm presenting good performance in the discrete action space and thus it was used in this study. The action of DQN agent is required to represent design variables of RC beam. However, the number of design variables of RC beam is too many to represent by the action of conventional DQN. To solve this problem, multi-agent DQN was used in this study. For more effective reinforcement learning process, DDQN (Double Q-Learning) that is an advanced version of a conventional DQN was employed. The multi-agent of DDQN was trained for optimal structural design of RC beam to satisfy American Concrete Institute (318) without any hand-labeled dataset. Five agents of DDQN provides actions for beam with, beam depth, main rebar size, number of main rebar, and shear stirrup size, respectively. Five agents of DDQN were trained for 10,000 episodes and the performance of the multi-agent of DDQN was evaluated with 100 test design cases. This study shows that the multi-agent DDQN algorithm can provide successfully structural design results of RC beam.

• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.12
• /
• pp.1073-1078
• /
• 2012

This paper deals with the trajectory planning of multi agent robots using complex potential theory for robot soccer. The complex potential theory is introduced, then the circle theorem is used to avoid obstacles, and the vortex pair is used to make precise kicking direction of robot. Various situations of robot soccer are simulated and the effect of vortex strength and the speed of robots are discussed and the better way to avoid obstacles and to kick the precise direction is found. The feasibilities of complex potential theory to apply for the multi agent robots are successful.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.879-884
• /
• 2004

This paper describes the development of a synthetic multi-agent called KMITL Cadence 2003. KMITL Cadence 2003 is a robotic soccer simulation team consisting of eleven autonomous software agents. Each agent operates in a physical soccer simulation model called Robocup Soccer Server which provides fully distributed and real-time multi-agent system environment. All teammates have to cooperate to achieve the common goal of winning the game. The simulation models many aspects of the football field such as noise in ball movements, noisy sensors, unreliable communication channel between teammates and actuators, limited physical abilities and restricted communication. This paper addresses the algorithm to develop the soccer agents to perform basic actions which are scoring, passing ball and blocking the opponents effectively. The result of this development is satisfactory because the successful scoring attempts is increased from 11.1% to 33.3%, successful passing ball attempts is increased from 22.08% to 63.64%, and also, successful intercepting attempts is increased from 88% to 97.73%.

• International Journal of Automotive Technology
• /
• v.7 no.2
• /
• pp.145-154
• /
• 2006

In the companion paper, the composition and structure of the MATDYMO (Multi-Agent for Traffic Simulation with Vehicle Dynamic Model) were proposed. MATDYMO consists of the road management system, the vehicle motion control system, the driver management system, and the integration control system. Among these systems, the road management system and the integration control system were discussed In the companion paper. In this paper, the vehicle motion control system and the driver management system are discussed. The driver management system constructs the driver agent capable of having different driving styles ranging from slow and careful driving to fast and aggressive driving through the yielding index and passing index. According to these indices, the agents pass or yield their lane for other vehicles; the driver management system constructs the vehicle agents capable of representing the physical vehicle itself. A vehicle agent shows its behavior according to its dynamic characteristics. The vehicle agent contains the nonlinear subcomponents of engine, torque converter, automatic transmission, and wheels. The simulation is conducted for an interrupted flow model and its results are verified by comparison with the results from a commercial software, TRANSYT-7F. The interrupted flow model simulation is implemented for three cases. The first case analyzes the agents' behaviors in the interrupted flow model and it confirms that the agent's behavior could characterize the diversity of human behavior and vehicle well through every rule and communication frameworks. The second case analyzes the traffic signals changed at different intervals and as the acceleration rate changed. The third case analyzes the effects of the traffic signals and traffic volume. The results of these analyses showed that the change of the traffic state was closely related with the vehicle acceleration rate, traffic volume, and the traffic signal interval between intersections. These simulations confirmed that MATDYMO can represent the real traffic condition of the interrupted flow model. At the current stage of development, MATDYMO shows great promise and has significant implications on future traffic state forecasting research.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1938-1943
• /
• 2005

Network human-robot interface is an important research topic. In home application, users access the robotic system directly via voice, gestures or through the network. Users explore a system by using the services provided by this system and to some extend users are enable to participate in a service as partners. A service may be provided by a robot, a group of robots or robots and other network connected systems (distributed sensors, information systems, etc). All these services are done in the network environment, where uncertainty such as the unstable network connection, the availability of the partners in a service, exists. Moreover, these services are controlled by several users, accessing at different time by different methods. Our research aimed at solving this problem to provide a high available level, flexible coordination system. In this paper, a multi-agent framework is proposed. This framework is validated by using our new concept of slave agents, a responsive multi-agent environment, a virtual directory facilitator (VDF), and a task allocation system using contract net protocol. Our system uses a mixed model between distributed and centralized model. It uses a centralized agent management system (AMS) to control the overall system. However, the partners and users may be distributed agents connected to the center through agent communication or centralized at the AMS container using the slave agents to represent the physical agents. The system is able to determine the task allocation for a group of robot working as a team to provide a service. A number of experiments have been conducted successfully in our lab environment using Issac robot, a PDA for user agent and a wireless network system, operated under our multi agent framework control. The experiments show that this framework works well and provides some advantages to existing systems.