Search

Along with rapid prevalence of Internet of things (IoT) and mobile network services, network function virtualization (NFV) based infrastructure requires dynamic computational resource adjustments in response to time-varying environments (network traffic, resource usage, failure status, etc). To provide agile resource control and adaptiveness, it is effective to predict a virtual server load by means of machine learning technologies for proactive control. In this paper, we propose a regression analysis model utilizing a sparse modeling to predict the average server load in a future specific time period on the basis of the server load in a past specific time period. We target at the number of access to a Web server as a virtual server load. The model can reduce the prediction error compared to a traditional least square method. Besides, we propose two methods to further reduce prediction errors: (A) an algorithm for determining the time period targeted for explanatory variables, and (B) two-stage regression analyses. As a result of MATLAB calculations, we show that application of the sparse modeling can reduce the prediction error by more than 40%, and besides, the above two proposed methods are effective at further reducing the prediction error with a few minutes’ learning time. Our model also contributes to make humans’ posterior analysis easier by making a short list of explanatory variables of data than conventional methods.

The Global Navigation Satellite System-Acoustic ranging combination technique (GNSS-A) has achieved many historical geoscientific observations—a coseismic huge deformation following the 2011 Tohoku-oki earthquake, an interseismic coupling condition along the Nankai Trough megathrust zone, and so on. However, even with the latest observation technology, it takes about 3 years to detect an interseismic small deformation. We then cannot detect a small coseismic event or a slow slip event. In order to exceed such detection limit, elimination of errors‐‐especially errors from marine fields—by analytical approaches is required. In the current analysis strategy, only an analyzing of the shallowest ocean structure with the most disturbances has been conducted. Here, we developed a new analytical strategy that can extract a gradient field of a deeper structure and applied it to actual data. Resultant extracted ocean fields were consistent with the JCOPE-2 ocean model and real ocean fields. Resultant time series of the seafloor crustal deformations were also refined by this new strategy. These total results show the validity of the new analysis technique and the extracted natural phenomenon.

This presentation introduces two experiments of steady-state responses utilization for the flight simulator trainings. In the first experiment, changes of auditory steady state response (ASSR) during a series of flight simulator training were assessed. In the on-going second experiment, we are now trying to capture the course of visual abnormality in high-g training via steady-state visually evoked potential (SSVEP).

Present programming education methods bring about many big gaps in understanding of many novice programmers. Novices suffering from big gaps in understanding would lose confidence to continue to learn programming. We introduce a method named AtoP to grain the big gaps into small gaps so that a novice can overcome the gaps. The method aims to associate programming with knowledge in many areas through interactive supervision and self-efficacy promotion.

This paper proposes a method to refine education contents with absolute measures. The measure is constructed based on the collective intelligence approach. The method values the motivation and the understanding of students. Operations, which are units of service in programming education courses are selected automatically to reduce refinement costs.

In wireless sensor networks (WSN), energy saving is the main issue of research, and many energy saving techniques have been proposed. Dynamic Voltage Scaling (DVS) and Dynamic Modulation Scaling (DMS) are techniques reduce energy consumption by using the idle time of the sensor node. When we use both techniques at the same time, an appropriate mechanism of cooperation is required because of the limitation of available idle time of the sensor node. Therefore, we propose a cooperative power. saving technique that applies DVS and DMS to sensor nodes for minimizing energy consumption. It uses a prediction mechanism that estimates the load of the processor and the radio communication device based on the log data to realize proper cooperation. We observed that the adaptation of the proposed technique is able to achieve the energy reduction up to 40 percent compared to without any energy saving techniques.

Recent years, the mobile ad hoc networks (MANETs) research community has been very active. Thus, it is of growing importance to build distributed network applications on top of MANETs. DHT is known as a popular technique to build efficient distributed applications in the Internet. However, in MANETs, DHT needs to consider network traffics, moving of nodes and energy efficiency more carefully than the Internet. We propose a technique for construction of overlay networks for MANETs in consideration for topologies of physical networks. Our technique realizes the pure P2P network that does not require special nodes like cluster heads or super peers. It gives a node ID to each node in a network using information from physically neighbor nodes so that the node has a node ID value close to that of neighbor nodes. Thus the neighbor nodes in the physical network will be also close in the logical network. When a node moves, a node ID will be reallocated to meet the above requirement. We evaluate the performance of the proposed technique by the simulations.

A video-based meeting support system is proposed. It aims to support daily group meetings without extra system operations. Purposes, usage and architecture of the system is described.