Itsuki Noda

The RoboCup Soccer Server and CMUnited: Implemented Infrastructure for MAS Research

The RoboCup Soccer Server and associated client code is a growing body of software infrastructure that enables a wide variety of multiagent systems research. The Soccer Server is a multiagent environment that supports 22 independent agents interacting in a complex, real-time environment. AI researchers have been using the Soccer Server to pursue research in a wide variety of areas, including real-time multiagent planning, real-time communication methods, collaborative sensing, and multiagent learning. This article describes the current Soccer Server and the champion CMUnited soccer-playing agents, both of which are publically available and used by a growing research community. It also describes the ongoing development of FUSS, a new, flexible simulation environment for multiagent research in a variety of multiagent domains.

Is Dial-a-Ride Bus Reasonable in Large Scale Towns? Evaluation of Usability of Dial-a-Ride Systems by Simulation

Dial-a-ride systems are attracting attention as a new style of transportation systems for urban areas. While it has been reported that such systems improve the usability of bus systems when applied in a small town or area, it is not obvious how and under what conditions the systems are effective in comparison to traditional fixed-route bus systems. We conducted two computer simulations of dial-a-ride and fixed-route systems in order to compare the usability and profitability of both systems. Simulation results indicated that: (1) The usability of the dial-a-ride system with a fixed number of buses drops very quickly when the number of requests(demands) increases. (2) When we increase the number of buses proportionally to the number of demand, the usability of the dial-a-ride system is improved more significantly than that of the fixed-route system. (3) When frequency of demands is sufficiently, the dial-a-ride system is a reasonable solution from the both usability and profitability perspectives.

RoboCup-Rescue: an international cooperative research project of robotics and AI for the disaster mitigation problem

This paper introduces the RoboCup-Rescue Simulation Project, a contribution to the disaster mitigation, search and rescue problem. A comprehensive urban disaster simulator is constructed on distributed computers. Heterogeneous intelligent agents such as fire fighters, victims and volunteers conduct search and rescue activities in this virtual disaster world. A real world interface integrates various sensor systems and controllers of infrastructures in the real cities with the real world. Real-time simulation is synchronized with actual disasters, computing complex relationship between various damage factors and agent behaviors. A mission-critical man-machine interface provides portability and robustness of disaster mitigation centers, and augmented-reality interfaces for rescue in real disasters. It also provides a virtual- reality training function for the public. This diverse spectrum of RoboCup-Rescue contributes to the creation of the safer social system.

Evaluation of decision support systems for emergency management

We propose a new simulation methodology in RoboCup Rescue framework. By involving humans in simulations, we can achieve more goals than those with only autonomous agents. We focus on the evaluation of decision support system (DSS) with the methodology and show how to design and evaluate DSS. By using the method we described in this paper, the effectiveness of DSS can be measured quantitatively.

Language Design for Rescue Agents

We are proposing a model of communication and a specification of a language for civilian agents in RoboCup Rescue Simulation System.Robust information systems are critical infrastructures for rescue activities in huge disasters. In order to simulate (and evaluate) a certain rescue information system, we need to design abstract model of agents communication, which is an important factor to affect the performance of the rescue activities. Especially communication among civilians, who are the majority in damaged area, will be the primary information source for rescue agents.In order to build the abstract model, we design four layers model of communication, which consists of knowledge, attention, device, and transmission layers. Using the model, we can discuss and implement uncertainty and effectiveness of various communication method including mobile phones, broadcasts, blackboards and so on.Then, we design specification languages for civilian agents behave in the simulated disaster world, which can reflect natural language features like uncertainty and lack of words.

Hidden Markov Modeling of Team-Play Synchronization

Imitation Learning is considered both as a method to acquire complex human and agent behaviors, and as a way to provide seeds for further learning. However, it is not clear what is a building block in imitation learning and what is the interface of blocks; therefore, it is difficult to apply imitation learning in a constructive way. This paper addresses agents’ intentions as the building block that abstracts local situations of the agent and proposes a hierarchical hidden Markov model (HMM) in order to tackle this issue. The key of the proposed model is introduction of gate probabilities that restrict transition among agents’ intentions according to others’ intentions. Using these probabilities, the framework can control transitions flexibly among basic behaviors in a cooperative behavior. A learning method for the framework can be derived based on Baum-Welch’s algorithm, which enables learning by observation of mentors’ demonstration. Imitation learning by the proposed method can generalize behaviors from even one demonstration, because the mentors’ behaviors are expressed as a distributed representation of a flow of likelihood in HMM.

Segmentation of environments using hidden Markov modeling of other agents

Learning by observation and imitation has been focused as a way for intelligent agents and robots to realize exible behaviors [1, 6]. One open issue on the imitation learning is segmentation: When a learner observes behaviors of others (demonstrators), it must estimate internal representation of sensed environment and demonstrators intentions, and segment the representation into parts that can be building blocks for constructing generalized imitation. The diÆculty of the segmentation of environment is that the segmentation should change according to agents current intention. This means that segmentation of environment should be acquired in the same time of learning intentions in the context of imitation learning.

Urban flood simulation as a component of integrated earthquake disaster simulation

Various reasons cause floods in cities, RoboCup Rescue simulation system has been designed so that various disasters can be plugged into the existing system. However, it is hard to handle water in RoboCup Rescue simulators. In this paper, we show flood simulation based on diffusion equation model is easily implemented using IDSS and IDSSs scalability is useful for flood simulations.

Hidden Markov Modeling of Multi-agent Systems and Its Learning Method

Two frameworks of hidden Markov modeling for multi-agent systems and its learning procedure are proposed. Although a couple of variations of HMMs have been proposed to model agents and their interactions, these models have not handled changes of environments, so that it is hard to simulate behaviors of agents that act in dynamic environments like soccer. The proposed frameworks enables HMMs to represent environments directly inside of state transitions. I first propose a model that handles the dynamics of the environments in the same state transition of the agent itself. In this model, the derived learning procedure can segment the environments according to the tasks and behaviors the agent is performing. I also investigate a more structured model in which the dynamics of the environments and agents are treated as separated state transitions and coupled each other. For this model, in order to reduce the number of parameters, I introduce symmetricity among agents. Furthermore, I discuss relation between reducing dependency in transitions and assumption of cooperative behaviors among multiple agents.

Communication in multi-agent simulation

In multi-agent simulations (MA-Sim), formalization of communication among agents is an important issue on building simulation systems. Agent communication has two significant features in MA-Sim: 1) it is a complex process that consists of various physical phenomena; and 2) it is an important functional operator that influences agents behaviors directly. In this paper, we discuss the modeling of the communication as complex physical phenomena in the simulation system. Moreover, we investigate the interaction between communications and agents behavior based on multi-scenario-based agent programming.