Koichi Kurumatani

Smooth traffic flow with a cooperative car navigation system

With maturation of ubiquitous computing technology, it has become feasible to design new systems to improve our urban life. In this paper, we introduce a new application for car navigation in a city. Every car navigation system in operation today has the current position of the vehicle, the destination, and the currently chosen route to the destination. If vehicles in a city could share this information, they could use traffic information to globally plan semi-optimal routes for each vehicle. Thus, we propose a cooperative car navigation system with route information sharing (RIS). In the RIS system, each vehicle transmits route information (current position, destination, and route to the destination) to a route information server, which estimates future traffic congestion using this information and feeds its estimate back to each vehicle. Each vehicle uses the estimation to re-plan their route. This cycle is then repeated. Our multiagent simulation confirmed the effectiveness of the proposed RIS system. The average travel time of drivers using the RIS system is substantially shorter than the time of drivers who chose shortest distance or simple shortest time estimates. Moreover, as the number of RIS users increases, the total amount of traffic congestion in the city decreases.

Modeling of Theme Park Problem with Multiagent for Mass User Support

We propose the Theme Park Problem as one example of test bed for mass user support. The objective of the problem’s solution is to coordinate many visiting agents’ behavior in a way that increases social welfare without reducing individual satisfaction. For the computer simulation, we design four types of basic coordination algorithms, and the simulation’s results are used to discuss the basic characteristics of the problem.

CONSORTS-S: A mobile sensing platform for context-aware services

This paper proposes a mobile sensing platform that provides context-aware services for mobile users by accessing surrounding wireless sensor networks. The platform provides the following facilities: 1) communicating to wireless sensor networks via a mobile sensor router attached to a userpsilas mobile phone, 2) analyzing the sensed data derived from networks by cooperating with sensor middleware on a remote server to capture onepsilas contexts, and 3) providing context-aware services for mobile users of cellular telephones. The platform is designed to process the sensed data effectively through cooperation between a mobile phone, a mobile sensor router, and a sensor middleware on a remote server. In this paper, we describe the concept of a mobile sensing platform: CONSORTS-S and its architecture, and show a healthcare service as a prototype application of CONSORTS-S. The healthcare service is aware of userspsila conditions such as heartbeat, posture, and movement through monitors of physiological signals (e.g., electrocardiograph, thermometer, and 3-axis accelerometer) and environmental conditions (e.g. room temperatures) by communication with environmental sensors.

U-Mart: A Test Bed for Interdisciplinary Research into Agent-Based Artificial Markets

U-Mart is a research project to construct a virtual stock market in which any preregistered agent can participate via networks. The aim of the project is to provide a forum where many agents of both computer programs and human traders participate in an auction and compete in a market, and also to provide a forum where research on market structure analysis is carried out. In this chapter, we introduce the concept and the system specification of U-Mart, and describe its importance and research possibilities in the context of economics, computer sciences, and artificial intelligence.

Multi-Agent for Mass User Support

Agent-based Methodology (ABM) is becoming indispensable for the inter-disciplinary study of social and economic complex adaptive systems. The essence of ABM lies in the notion of autonomous agents whose behavior may evolve endogenously and can generate and mimic the corresponding complex system dynamics that the ABM is studying. Over the past decade, many computational intelligence (CI) methods have been applied to the design of autonomous agents, in particular, their adaptive scheme. This design issue is non-trivial since the chosen adaptive schemes usually have great impact on the generated system dynamics. Robert Lucas, one of the most influential modern economic theorists, has suggested using laboratories with human agents, also known as Experimental Economics, to help solving the design issue. While this is a promising approach, laboratories used in the current experimental economics is not computationally equipped to meet the demands of the task. This paper attempts to materialize Lucas’ suggestion by establishing a laboratory where human subjects are equipped with the computational power that satisfies the computational equivalence conditions.

Distributed visitors coordination system in theme park problem

A distributed visitors coordination system is proposed as an application of a massively multi-agent system. In the system, some agents register their next destination using an information device such as a cellular phone, and this information is used to reduce the effect of the time delay between decision-making and effect-emergence. This delay causes queue lengths to oscillate. However, it is troublesome for agents to continuously register their next destination. To compensate them, exclusive queues are made available to agents registering their next destination. Computer simulation of the theme park problem, showed that when all agents avoid the congestion by registering their next destination, the total waiting time is minimized and queue length oscillation is reduced.

CONSORTS: A Multiagent Architecture for Service Coordination in Ubiquitous Computing

One of the fundamental issues of Ubiquitous Computing is concerned with the coordination gaps between devices, services, and users. When numerous devices, various information services, and users who have different intentions are physically co-located in a environment, how can we coordinate the services and devices to assist a particular user in receiving a particular service so as to maximize the user’s’ satisfaction there? In order to solve this human-centered service coordination issue, we have been developing a multiagent architecture for Ubiquitous Computing, called CONSORTS (Coordination System of Real world Transaction Services). In this paper, we first outline some coordination gaps in Ubiquitous Computing, and describe three design concepts of the CONSORTS to bridge the gaps:physically-grounding, cognitive resource managements, andlocation-mediated service coordination. Then, we show the outline of CONSORTS architecture, and two applications of CONSORTS, context-aware information assist systems in museums and wireless-LAN based location system on AgentCities Networks.

ZigBee based indoor localization with particle filter estimation

Highly-functional mobile devices, such a smart phone, have appeared. Location based services of the mobile devices are assimilated in a variety of ways. Then, Indoor localization sensor is necessary to access the location based services seamlessly. This paper researched the performance of indoor localization with ZigBee based particle filter. This paper showed this method can localize a resting target within 2.0 meter accuracy and can localize a moving target with an area levels localization. ZigBee based particle filter can be one of some options for location based services in indoor spaces.

Spatio-temporal sensor data management for context-aware services: designing sensor-event driven service coordination middleware

How various kinds of sensor devices are handled, and how numerous lower-level sensor data are managed and integrated into higher-level context representations are important issues to realize context-aware services. We have been developing Sensor-Event-Driven Service Coordination Middleware (SENSORD) to fill coordination gaps between higher-level services and lower-level sensors. The SENSORD system obtains and stores sensor data into an in-memory data container to achieve fast, complex analysis of the sensed data. The facility of real-time analysis, which includes periodical evaluations of spatio-temporal conditions, enables application developers to outsource context-aware mechanisms. As described in this paper, we first examine middleware of context-aware services and our approaches. We then show the outline of SENSORD and an exemplary application: an indoor emergency response system in our laboratories. Preparing for emergency situations, such as fire emergencies, it manages sensor devices (thermometers, hygrometers, vision systems, microphone arrays, etc.) to detect emergent situations.

ComPass system: an low power wireless sensor network system and its application to indoor positioning

In this paper, we propose a wireless sensor network system called ComPass System, which uses low-power radio to provide services in everyday environments, such as indoor positioning services in public spaces. Although such environments present difficult hindrances such as dynamic changes of physical status, the ComPass System is designed to operate in such environments. The ComPass System 1) uses low-power radio and VHF radio band, 2) has low power consumption and functionality with cell batteries, and 3) has higher-level function such as bit-rate detection and signal correction on the ComPass Node----a node device designed for the ComPass System. We constructed an indoor autonomous positioning system in a commercial building, Yokohama Landmark Plaza, Minato-Mirai area, Japan, based on the ComPass System