• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.2036-2041
• /
• 2005

The underwater multi-agent technology has many potential for the various activities related to ocean development/conservation in the near future. For example, in such fields as water pollution investigation, aquaculture control, or coral reef research, we feel a growing need for a system that realizes underwater continuous monitoring in the wide rang e. In this case, the target monitoring area will be sliced planar hierarchically toward the depth as monitoring layers, and many AUVs arranged on each layer track the given trajectory and gather various environmental information continuously, with communicating each other in the layer or with other layers. To realize those systems we need to develop AUV multi-agent technologies. So we are now building basic systems for basin experiment for the development of AUV multi-agent behavior. We must experience many situations and problems to be solved for the development of its elemental technologies by using real systems as well as our computer simulations. In this paper we introduce our concept of the experiment in the near future and the hardware/software design of our two types of handy AUVs and ultrasound ranging/communication system for that experiment. One AUV is designed using a 17inches-diameter glass sphere with DOS/V and RT-Linux based subsystems, which is intended to use not only in the basin but also in the calm real sea. The other AUV is designed for the basin experiment using a 7inches-diameter acrylic sphere with low-cost embedded system with SH-2 based subsystems. The basin experiment to verify the basic AUV facilities and ultrasound ranging for position detection was carried out.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.1
• /
• pp.11-19
• /
• 2024

Multi-agent systems can be utilized in various real-world cooperative environments such as battlefield engagements and unmanned transport vehicles. In the context of battlefield engagements, where dense reward design faces challenges due to limited domain knowledge, it is crucial to consider situations that are learned through explicit sparse rewards. This paper explores the collaborative potential among allied agents in a battlefield scenario. Utilizing the Multi-Robot Warehouse Environment(RWARE) as a sparse reward environment, we define analogous problems and establish evaluation criteria. Constructing a learning environment with the QMIX algorithm from the reinforcement learning library Ray RLlib, we enhance the Agent Network of QMIX and integrate Random Network Distillation(RND). This enables the extraction of patterns and temporal features from partial observations of agents, confirming the potential for improving the acquisition of sparse reward experiences through intrinsic rewards.

• International Journal of Automotive Technology
• /
• v.7 no.1
• /
• pp.25-34
• /
• 2006

For decades, simulation technique has been well validated in areas such as computer and communication systems. Recently, the technique has been much used in the area of transportation and traffic forecasting. Several methods have been proposed for investigating complex traffic flows. However, the dynamics of vehicles and diversities of driver characteristics have never been considered sufficiently in these methods, although they are considered important factors in traffic flow analysis. In this paper, we propose a traffic simulation tool called Multi-Agent for Traffic Simulation with Vehicle Dynamics Model (MATDYMO). Road transport consultants, traffic engineers and urban traffic control center managers are expected to use MATDYMO to efficiently simulate traffic flow. MATDYMO has four sub systems: the road management system, the vehicle motion control system, the driver management system, and the integration control system. The road management system simulates traffic flow for various traffic environments (e.g., multi-lane roads, nodes, virtual lanes, and signals); the vehicle motion control system constructs the vehicle agent by using various vehicle dynamic models; the driver management system constructs the driver agent capable of having different driving styles; and lastly, the integrated control system regulates the MATDYMO as a whole and observes the agents running in the system. The vehicle motion control system and driver management system are described in the companion paper. An interrupted and uninterrupted flow model were simulated, and the simulation results were verified by comparing them with the results from a commercial software, TRANSYT-7F. The simulation result of the uninterrupted flow model showed that the driver agent displayed human-like behavior ranging from slow and careful driving to fast and aggressive driving. The simulation of the interrupted flow model was implemented as two cases. The first case analyzed traffic flow as the traffic signals changed at different intervals and as the turning traffic volume changed. Second case analyzed the traffic flow as the traffic signals changed at different intervals and as the road length changed. The simulation results of the interrupted flow model showed that the close relationship between traffic state change and traffic signal interval.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.01a
• /
• pp.156-163
• /
• 2001

Business forecasting is vital to the success of business. There has been an increasing demand for building business forecasting software system to assist human being to do forecasting. However, the uncertain and complex nature makes is a challenging work to analyze, design and implement software solutions for business forecasting. Traditional forecasting systems in which their models are trained based on small collection of historical data could not meet such challenges at the information explosion over the Internet. This paper presents an agent oriented business forecasting approach for building intelligent business forecasting software systems with high reusability. Although agents have been applied successfully to many application domains. little work has been reported to use the emerging agent oriented technology of this paper is that it explores how agent can be used to help human to manage various business forecasting processes in the whole business forecasting life cycle.

• Journal of Information Processing Systems
• /
• v.8 no.3
• /
• pp.471-482
• /
• 2012

For disaster exploration and surveillance application, this paper aims to present a novel application of a multi-robot agent based on WSN and to evaluate a multi-hop communication caused by the robotics correspondingly, which are used in the uncertain and unknown subterranean tunnel. A Primary-Scout Multi-Robot System (PS-MRS) was proposed. A chain topology in a subterranean environment was implemented using a trimmed ZigBee2006 protocol stack to build the multi-hop communication network. The ZigBee IC-CC2530 modular circuit was adapted by mounting it on the PS-MRS. A physical experiment based on the strategy of PS-MRS was used in this paper to evaluate the efficiency of multi-hop communication and to realize the delivery of data packets in an unknown and uncertain underground laboratory environment.

• 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.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2001.01a
• /
• pp.73-77
• /
• 2001

In this paper, we propose a virtual action system based on information agents, We call the system the MultiHammer, MultiHammer can be used for studying and analyzing online actions. MuiltiHammer provides functions of implement-ing a meta online action site and an experiment environ-ment. We have been using MultiHammer as an experiment as an experiment environment for BiddinBot. BiddingBot aims at assisting users to bid simultaneously in multiple online auctions. In order to bid simultaneously in multiple online auctions. In order to analyze the behavior of BiddngBot, we need to pur-chase a lot of items. It is hard for us to prepare a lot of fund to show usability and advantage of BiddingBot. MultiHam-mer enables us to effectively analyze the behavior of BiddingBot. MultiHammer consists of three types of agents for information collecting data storing and auctioning. Agents for information wrappers. To make agent work as wrarp-pers, we heed to realize software modules for each online action site. Implementing these modules reguires a lot of time and patience. To address this problem, we designed a support mechanism for developing the modules. Agents for data storing record the data gathered by agents for informa-tion collecting. Agents for auctioning provide online services using data recorded by agents for data storing. By recording the activities in auction sites. MultiHammer can recreate any situation and trace auction for experimentation, Users can participate in virtual using the same information in real online auctions. Users also participate in real auc-tions via wrapper agents for information collecting

• Journal of the Korea Institute of Military Science and Technology
• /
• v.26 no.4
• /
• pp.361-370
• /
• 2023

Swarm has recently become a critical component of offensive and defensive systems. Multi-agent reinforcement learning(MARL) empowers swarm systems to handle a wide range of scenarios. However, the main challenge lies in MARL's scalability issue - as the number of agents increases, the performance of the learning decreases. In this study, transfer learning is applied to advanced MARL algorithm to resolve the scalability issue. Validation results show that the training efficiency has significantly improved, reducing computational time by 31 %.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.14 no.4
• /
• pp.284-291
• /
• 2021

In this paper, we study the consensus problem for multi-agent systems with input saturations. The goal of consensus is to achieve a swarming behavior of multi-agent systems by reaching the agreement through information exchange. This paper considers agents modeled by first-order dynamics with input saturations. In order to guarantee the global convergence of the agents, it is assumed that the agents are stable. Moreover, considering the disturbances, we propose the PI based duplex control method to achieve the consensus. The proposed P controller and I controller are composed of different information network. Then, we investigate the conditions of the information networks and the control gains of P, I controllers to achieve the consensus applying the Lyapunov stability theorem and the Lasalle's Invariance Principle. Finally, we conduct the simulations to validate the theoretical results.

• Journal of Institute of Control, Robotics and Systems
• /
• v.19 no.5
• /
• pp.390-397
• /
• 2013

This paper presents a consensus algorithm for uMAS (uncertain Multi-Agent Systems). Unlike previous results in which only nominal models for agents are considered, it is assumed that the uncertain agent model belongs to a known polytope set. In the middle of deriving the proposed algorithm, a convex set is found which includes all uncertainties in the problem using convexity of the polytope set. This set plays an important role in designing the consensus algorithm for uMAS. Based on the set, a consensus condition for uMAS is proposed and the corresponding consensus design problem is solved using LMI (Linear Matrix Inequality). Simulation result shows that the proposed consensus algorithm successfully leads to consensus of the state of uMAS.