西出 亮

Our research group is working on soybeans which the quantity of yielding is difficult to predict. We focus on the common characteristics observed at multiple cultivation points, in order to examine methods to acquire new knowledge in deciding the work based on the amount of yields. Our previous study has examined a method to discover optimal patterns using qualitative value of

As mobile device technology advances, content delivery among nearby devices becomes increasingly important at such places as where people gather face-to-face for seminars or meetings. Conventional ways such as cloud-based services, ad-hoc contents dissemination methods and file sharing peripheral devices are not enough to support this style of content delivery. In this paper, we propose a two-layered content delivery using multihop wireless communication built on an ad-hoc network. This network supports content delivery among tens of nearby devices even though the underlying ad-hoc network has limitation of the number of devices. We also show experiment results to reveal how the contents delivery proceeds in the network and how the network topology affects the delivery time. Important findings in the experiments are that 1MB data can be delivered to 16 devices in approximately 5.25 seconds using Wi-Fi based ad-hoc connection on iOS devices, and that reducing the number of devices simultaneously sending the data leads to the less required time for delivery.

As mobile terminals become popular, more people are getting involved with collaboratively searching Web sites in order to achieve a common purpose with others. When performing such a task, we want to save the usage of the Internet traffic as much as possible because the transfer speed of mobile terminals may be limited when the data transfer usage exceeds a certain amount of gigabytes. To reduce the Internet traffic, we build a proxy system with the peer cache mechanism to share the Web contents stored on participating mobile terminals, focusing on the existence of the Web contents which are accessed multiple times from different terminals. Our experimental results reveal that about 20% of the Internet traffic has been reduced when four people engaged in a collaborative Web search task to find good restaurants for a year-end party.

Purpose-The purpose of this study is to reduce internet traffic when performing collaborative Web search. Mobile terminals are now in widespread use and people are increasingly using them for collaborative Web search to achieve a common goal. When performing such searches, the authors want to reduce internet traffic as much as possible, for example, to avoid bandwidth throttling that occurs when data usage exceeds a certain quota. Design/methodology/approach-To reduce internet traffic, the authors use a proxy system based on the peer cache mechanism. The proxy shares Web content stored on mobile terminals participating in an ad hoc Bluetooth network, focusing on content that is accessed multiple times from different terminals. Evaluation of the proxy's effectiveness was measured using experiments designed to replicate realistic usage scenarios. Findings-Experimental results show that the proxy reduces internet traffic by approximately 20 per cent when four people collaboratively search the Web to find good restaurants for a social event. Originality/value-Unlike previous work on co-operative Web proxies, the authors study a form of collaborative Web caching between mobile devices within an ad hoc Bluetooth network created specifically for the purpose of sharing cached content, acting orthogonally to (and independently of) traditional hierarchical Web caching.

Maximizing the quality of service experienced by passengers of public transportation systems is of great importance to service operators. Accurate and timely information about passenger load, both at stops and on board vehicles, can help in many ways to improve the quality of service. This information is currently not readily available using data collection techniques that are transparent to passengers. We describe a method and feasibility study for estimating transit passenger population at two bus stops by passively monitoring Wi-Fi network activity of mobile devices. The collected data is filtered and analyzed to distinguish passengers' mobile devices from non-passenger devices. Our experimental results show that filtered Wi-Fi activity correlates well with observed passenger population at bus stops and has the potential to provide near real-time information.

We present a method to form a co-located group with detection of relative position using camera image, for enabling a user to share digital content instantly. The method uses only incorporated components on a mobile device without using peripheral devices such as RGB-D camera and independent of work environment or arrangement of users. Location of each user is detected based on the value of the on-device accelerometer and gyroscope. This provides a user highly co-located interaction anywhere. We also implement a collaborative Web search tool to which method is applied for supporting collaborative work through Web search.

Influence Maximization is the problem of choosing a small set of seed users within a larger social network in order to maximize the spread of influence under certain diffusion models. The problem has been widely studied and several solutions have been proposed. Previous work has concentrated on positive relationships between users, with little attention given to the effect of negative relationships of users and the corresponding spread of negative opinion. In this paper we study influence maximization in signed social networks and propose a new diffusion model called LT-S, which is an extension to the classical linear threshold model incorporating both positive and negative opinions. To the best of our knowledge, we are the first to study the influence maximization problem in signed social networks with opinion formation. We prove that the influence spread function under the LT-S model is neither monotone nor submodular and propose an improved R-Greedy algorithm called RLP. Extensive experiments conducted on real signed social network datasets demonstrate that our algorithm outperforms the baseline algorithms in terms of efficiency and effectiveness.

Uneven loading of busses degrades passengers' travel experience for a variety of reasons. Load balancing for busses requires information about the load, both the potential passenger load at the bus stop and the actual load currently aboard travelling busses. This paper describes a feasibility study for measuring bus passenger loads by detecting the periodic network probing activity from WiFi devices built into 'smart phones'. Our experimental results show that WiFi activity does correlate with observed passenger flow at bus stops and the load aboard a bus while en route. (C) 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

本稿では，平面上に存在するノード間を低ホップ数で結ぶSkip Geo Network（SGN) とその分散構成アルゴリズムを提案する．SGN は，Skip Graph を用いた論理ネットワークとノードの位置関係により構成されるドロネーネットワークのハイブリッド構造を持ち，各ノードはより近いノードから遠方ノードへのリンクを持つことができる．SGN を用いて，相手ノードを指定した通信に加えて，目的位置や領域を指定したジオキャスト（Geocast）を行うことができる．SGN は，ノード間でP2P型通信によりアドホックにネットワークを構成できるため，移動体通信，緊急時の情報散布などへ応用することができる．最後に，提案手法の有効性を確認するために数値シミュレーションにより，SGN のネットワークの特性とジオキャストの性能などについて検証する．In this paper，we propose Skip Geo Network(SGN), which is a scalable geographic network that connects computational entities or nodes in a 2D space by low P2P hop counts. SGN has a hybrid structure of logical networks and Delaunay networks generated using geographical coordinates where nodes reside. In SGN, each node has a long routing paths compared to method without SGN, which can enable Geocast to specified points or regions with low hop counts. SGN can be generated by P2P communication among nodes in ad-hoc network, and we expect that it is applicable to location-oriented systems, such as mobile hosts in MANET. We propose an autonomous generation algorithm of SGN and evaluate network parameters and geocast parameters of SGN by numerical simulations.

Researches on computing an optimal route according to user's preference or spatial contexts have been gaining focus recently, as car navigation and pedestrian navigation systems are getting accustomed to be used in daily lives. In this paper, the authors propose a method to compute a path to pass through as man), context areas as possible along its route, according to the user's needs. Specifically, the authors apply the concept to construct a Delaunay Diagram from neighbor nodes in Voronoi Diagram, and use Network Voronoi Diagram to construct a subgraph based on contexts. For route search, the authors use A * shortest path algorithm over the subgraph. To verify the efficiency of the proposed method with subgraphs, the authors perform route, search over subgraphs and base graph, which was generated from the road network of the real space, and compare the results. The results show that the route search by the proposed method consist of more paths with context areas than those performed over the base graph. Moreover the proposed method reduces the number of steps for path search.

This paper proposes a method for extracting the ROI using geometric computations, to obtain information of commonly-interested objects by multiple users. The authors use P2P Delaunay Network to share the regional information shot by cameras with peers, and extract the most intersecting area of camera ranges as ROI. In P2P Delaunay Network, nodes and edges refer to as computers and their connections. This proposed method can construct P2P Delaunay Network using locations of mobile devices as GPS phones or PDAs equipped with a camera, and process computations distributively with these devices. Thereby, a P2P Delaunay network can be constituted, without using the computer resource of other settlements. The method first partitions the plane and assigns the partitions to nodes according to their locations using P2P Delaunay Network. Then, each camera propagates its camera range information to nodes, and each node thus processes computation distributively. The node which has received multiple camera range information computes the intersecting area within its Voronoi region. Finally, the ROI can be determined, by computing the most intersecting area among all nodes. The authors examine the proposed method by evaluating the frequency of region partitioning and data transfer for each node through numerical simulation.

本稿では，スケーラブルなネットワーク基盤として，計算幾何の分野で知られるドロネー図をトポロジとして持つP2P ドロネーネットワークとその自律分散生成アルゴリズムを提案する．ここでは，まず本P2P ドロネーネットワークの特徴について述べ，次に生成アルゴリズムについて述べる．提案アルゴリズムは，各ノードの局所的な動きから，ノード間の幾何学的な位置関係を利用して接続関係を更新し続けるアルゴリズムであり，さらにノードが相互に情報交換することでネットワークを構成できる特徴を持つ．また，ノードが幾何学的退化状態にある場合も動作できる．本アルゴリズムにより，与えられた2 つのP2P ドロネーネットワークを融合することも可能であり，P2P パラダイムの持つスケーラビリティを活かしながら，システムの対象空間を段階的に拡張できる．提案アルゴリズムでは，ノードの3 つの操作を定義している．すなわち，局所ドロネー化操作と三角化通知操作が，局所的なドロネー図を自律分散的に生成し，委譲操作により，ノード間でノード情報の情報交換を行う．最後に，数値シミュレーションにより，P2P ドロネーネットワークの形成過程を，ノードへの負荷，P2P ドロネーネットワークへの収束ステップ数，各ノードの次数の変化，ネットワーク負荷などから検証し，提案アルゴリズムの有効性を確認する．また，本アルゴリズムの適用性についても議論する．This paper proposes a P2P Delaunay network whose topology is a Delaunay diagram wellknown in computational geometry as a scalable network infrastructure for spatial data management. We first discuss its features as a P2P network, and propose an algorithm to construct the network autonomously and distributively in P2P settings. In the proposed algorithm, nodes update their connection defined by node adjacency with respect to geometric location and generate local Delaunay networks of neighboring nodes, while they exchange node-location information to generate a network. The algorithm also works for the case nodes locate in geometrical degeneracy. Furthermore, the algorithm can also be applied to merging two independent P2P Delaunay networks. Owing to the algorithm, we can manage large target spaces using the P2P paradigm, and furthermore extend the target space incrementally y utilizing scalability that the P2P paradigm possesses. By numerical simulations, he authors examine the construction process of P2P Delaunay networks in terms of loads of odes, time-steps consumed for constructing P2P Delaunay networks, degree of each node, nd networkload cost. The applicability of the proposed algorithm for P2P models is also iscussed.

This paper proposes an interactive collaboration media in P2P settings, and discusses its behaviors using numerical simulations. Our aim is to construct this media system using locational data which avails for massive users to use in ad-hoc network environment. Users can read and write in this data sharing space to support collaborative interaction between users. We examine the system by simulation considering on the applications for school education, such as field studies or excursions. From the results, we show our system can realize load balancing of managed data to users. We believe that our system can enable scalable and dependable collaboration system holding massive users.

2005/

Recently, the remarkable advance on mobile terminal technologies has enabled us to develop full-scaled navigation systems for pedestrians. In most systems developed so far, research has been focused on visualization effect on maps to users to guide pedestrians to their destination smoothly. The authors here present a context-adaptable pedestrian navigation system, which provides users with the preferable route, considering the conditions of each area and the objectives of the users. A graph-based approach utilizing metadata is employed to represent logical space built over the real space, and a scoring scheme is provided by partitioning target region to its edges for approximation graph model for real space, which enables to expand target region. This scheme is dependent on context to find paths to pedestrian's destination. The usability of this system in many environments is examined by numerical simulations. The authors also discuss the possibilities of the system in practical settings in an open area, temporarily built-up sites.