With the deployment of smart homes, we find that human activity recognition (HAR) is essentially important to many applications, e.g., child/senior care, intelligent information push and exercise promotion. Although it is always better to build HAR model for each smart home to resolve the practical problem that homes have different floorplans or adopted sensors, it is intractable to acquire labeled data for each home due to cost and privacy. We thus propose a method to transfer the HAR model from multiple labeled source homes to the unlabeled target home. Specifically, we first generate transferable representations for the sensors of these homes, based on which we build the HAR model using the data of labeled source homes. Then, we employ the built HAR model into the unlabeled target home. Experiment results on CASAS dataset illustrate that our proposed method outperforms baseline methods in general and also avoids potential negative transfer caused by using only one source home.

The landscape of power grids is gradually changing. The growing number of electrical appliances as well as the outbreak of Electric Vehicles (EVs) is increasing the need for electricity. As a consequence, high-accuracy consumption predictions are necessary in order to both schedule and plan production and operations accordingly. The emergence of connected “tracking” devices opens up new data-sets into both Internet-of-Things and Big Data worlds. It provides information on human dynamics (people mobility behavior) and with it several opportunities. Electricity consumption is impacted by people movements as while moving they are not “connected” to the power grid. Therefore, predicting such movement patterns and volume could help electricity providers improve their own consumption predictions. This paper presents a system and methods used to predict people movement behaviors as well as detect any anomaly. A scoring system is used to both evaluate the dynamics predictions and raise alerts when the computed score surpasses established thresholds. This proposal is tested over a scenario using available datasets and demonstrates that modifications in people movement behavior is affecting the consumption profile.

Electrical devices which are controlled remotely by a smartphone have recently been spreading. With some typical existing systems, a user needs to identify a target device to control with only its less user-friendly ID. A system for association between actual position of the device and its position on smartphone display would allow the user to identify it easily. We call such a system “Smart Interaction System”. One of the key technologies for Smart Interaction System is localization of the device. Localization technique using some anchors with the given position to measure received signal strength indication (RSSI) of a device have been developed. They incur user burden to deploy the anchors and measure their positions. Thus alleviating the burden is still a technical challenge. As a solution to the challenge, this paper proposes a new 3D localization technique with a mobile robot including a floor cleaning robot. The mobile robot would reduce the burden by measuring RSSI while cleaning floor instead of the anchors. The proposed localization technique has been implemented with a floor cleaning robot. Experimental results show that the proposed localization technique provides the position of the device with approximately 1000mm of estimation error and is useful for Smart Interaction System.

How can you specify a device in the case of file sharing among multiple devices on that spot (e.g., a meeting picture of a social gathering or an alumni association)? One solution is to specify the partner device address (e.g., e-mail address or Bluetooth device address) on your own, but users sometimes feel annoyed or make a mistake with this operation. In this paper, we newly propose a device specification method by user operation of describing a circle in the air, enabling the user to be unaware of the device address. The proposed method uses the Doppler Effect of acoustic waves generated by the operation, enabling to specify multiple partner devices simultaneously and distinguish the device with which you do not want to share, without any pre-planned infrastructure. We measure its performance via an implemented application. The experimental results show that its performance is sufficient for the above use case.

Dual-mode cellular phones, including smartphones, are recently attracting much attention from both users and telecom operators throughout many countries. To avoid traffic concentration generated by these devices on 3G networks, it is important to lead users to connect with fixed networks via WLAN. However, keeping both 3G and WLAN interfaces on consumes a great deal of power, which considerably shortens the standby time. To achieve longer battery life, we propose a power-saving standby method (PSSM) whose features are: (1) constant deactivation of the WLAN interface except for when required; (2) no additional functions in a mobile operator's network; and (3) avoidance of a WLAN AP search distant from a cellular phone. We implemented PSSM as a BREW application and measured the electrical current in a cellular phone as well as the delay in email retrieval transaction. Experimental results using the prototype system show that PSSM achieves about 73%-134% longer battery life than a conventional dual-mode cellular phone, while PSSM keeps the application transaction delay short by reducing unproductive WLAN AP searches.

Recently, the rapid growth of Internet traffic has increased the volume of IT equipment in use. Reducing its power consumption is important to support such growing traffic. Furthermore, the energy consumption of servers and network equipment such as routers and switches accounts for about half the total energy consumption of all IT equipment. This is an incentive to reduce energy consumption at data centers, where numerous servers and network equipment are in operation. In this paper, we propose a new eco-friendly operation scheme for reducing the energy consumption of a data center that is effective in both virtualized and non-virtualized environments. In our proposal, we introduce a management server to monitor the server resource utilization periodically. In addition, the management server may dynamically change the data center configuration based on such utilization. For example, when the utilization is low, the management server changes the configuration so that the services can be provided via fewer servers and less network equipment, and turns off unnecessary servers and network equipment. We implement a system based on the proposed scheme, and evaluate its performance. The result shows that the proposed scheme is able to reduce the energy consumption of servers and network equipment by about 32.1% at maximum.

The greatest bottleneck in the updating of software programs (i.e. patching) is the time required to test a patched subset of systems for stability before rolling out the patch to all systems. For web servers, comprehensive tests are required for all web content. Even if we knew which modules would be updated by a patch, these tests would be unavoidable because we could not necessarily know the relationship between the web contents and web server modules. To reduce the test operation steps to perform, we propose a method of identifying the specific subset of web content and the web server modules that might be affected by a patch by identifying the relationship between the contents and modules. To estimate how much of a reduction could be made in the test operation steps, we conducted an evaluation using an actual web server system, which provides six web content groups. Our method identified that 23% of the web server modules were used by only one web content group. If we assume a patch would replaces these modules, the expected reduction in the test steps was 83% of the steps required to test all web content groups. The evaluation results implied that our method was effective in reducing the number of test steps.

In this poster, we propose a new tool, i.e. a node placement adviser, which supports deployment workers of sensor networks. The adviser tool is realized in the form of mobile terminal software and acts as a proxy of an additional node, as well as communicating with those already placed. Based on the communication results, it indicates how favorable the current location is to a worker.

The growth of Internet traffic has led to an increase in energy consumption by network equipment such as routers and switches. Consequently, energy consumption is becoming a key environmental, social, political and cost issue. Focusing on energy consumption by Internet Service Provider networks, the energy bottleneck is routers. Thus, it is imperative for us to reduce the energy consumption of routers. In this paper, we propose a new ECO-friendly distributed routing protocol (ECO-RP) based on OSPF. In our proposal, an ECO-RP entity periodically checks the amount of traffic forwarded by the ECO-RP entity. Moreover, an ECO-RP entity floods information on the amount of traffic to other entities such as the OSPF Link State Advertisement. This allows each ECO-RP entity to obtain information on traffic in the network, and to dynamically change OSPF link weights based on the information. For example, when the amount of overall network traffic is small, the entity changes link weights so that the traffic can be routed only through a subset of routers, and unneeded routers can shift into sleep mode. Several simulation studies, which assume two ECO-RP enabled networks, are conducted. The result shows that the proposed protocol is able to reduce energy consumption by about 18.5% at maximum without network congestion.

Prompt updating against the latest security threats is important, and conveniently achieved by using patches (i.e., software updates) on both servers and PCs in a timely manner. To apply patches safely, however, their impact should be examined in advance. Administrators need to pinpoint all the software programs that are potentially influenced by a patch. Our patch impact analyzer monitors the program's runtime behavior (i.e., library loadings) and identifies the full paths of those programs based on the kernel tracing method. Our method can identify even short-lived processes by introducing a self-constructed database of a program's full paths and process information. Evaluation using our prototype implementation and method shows that the full paths can be identified in about 99% of the library loading events.

Dual-mode devices, that have multiple network interfaces, have recently attracted both users and operators. Users are enabled to enjoy rich contents via high-speed and low-cost WLAN access. On the other hand, data offloading to WLAN network reduces operators' equipment cost for network infrastructure. However, a WLAN interface consumes much power even in the power save mode so that dual-mode devices would be totally unsatisfactory to users. This paper presents a consolidated solution available to many push applications to achieve “deep-sleep mode” in which WLAN interface is switched off as much as possible if not used. Application servers notify the dual-mode device of the event generation (e.g. receiving an incoming call, arriving an email, etc.). Receiving the notification, the dual-mode device activates WLAN interface and communicates with the application servers. We also implemented a prototype system, and experimental results show our method can reduce total power consumption.

Nowadays, many enterprises want to deploy Managed Network Services (MNS) so that their corporate resources are dedicated to their core businesses. Configuration management of network elements (NE) is a prerequisite of other parts of MNS, though it requires onsite personnel and is labor-intensive, so those systems that remotely and automatically manage NE configurations in behalf of MNS Providers (MNSPs)1 onsite personnel can help reduce service cost. However, the existing systems for this purpose, in which a monolithic server daemon serves for the entire customer base, cannot necessarily ensure independence of the service continuity for each customer, because an unexpected halt of service to a customer can bring the entire system down with it. We offer a customer equipment configuration management system (CECM), which allows MNSPs to remotely and automatically manage configurations of thousands of NEs geographically scattered in many sites simultaneously, while ensuring the independence of the service continuity CECM achieves this by adopting a novel Process-Per-Service (PPS) architecture, in which a set of independent system processes is exclusively used for the management of a customer's NE configurations. We also implement a prototype system to show the worst-case performance of CECM. The result shows that the configuration of 1000 emulated NEs for 100 customers can be simultaneously set within 45 seconds. While these results are preliminary, they suggest that CECM can help reduce service cost of providing MNSs.

Most sensor nodes usually have the ability to vary their own wireless transmission power. Hence they have the potential to control the network topology, which is the pattern of links connecting pairs of nodes in a network. In this paper, we propose a connectivity ascertainment based distributed topology control algorithm. The algorithm draws out the potential from nodes for constructing an energy-efficient network topology. Each node performs the algorithm and makes local decisions about its transmission power; these local decisions minimize the total transmission power of the network, while maintaining network connectivity. Specifically, each node ascertains the existence of an actual path that is a substitute for a wireless link to an own edge node, which requires the maximum transmission power among its own neighbors. To contribute to path ascertaining, the other nodes use locally-collected information. If a node can ascertain the existence, it decides to remove the wireless link to the edge node by reducing the transmission power. The node targets a new edge node from among its own current neighbors and performs the same process until there are no substitute paths to the target. We show that these local decisions collectively guarantee global connectivity and minimize the total transmission power of the network. We also evaluated the performance of our algorithm on terms of network connectivity and transmission power reduction using simulation studies.

Network monitoring is one of the most significant functions in network management to understand the state of a network in real-time. In network management, such as SNMP (simple network management protocol), polling is used for this purpose. If the time interval for two consecutive polling requests is too long, then we cannot understand the state of the network in real-time. Conversely, if it is too short, then the polling message traffic increases and imposes a heavy load on the network although we can understand the state of the network in real-time. Many schemes have been proposed for controlling the overheads of dynamic polling by throttling polling rates. Unfortunately, by only considering overheads, these schemes fail to take into account the rate at which information must be obtained in order to achieve management tasks. Examples of these tasks include checking for threshold violations and determining if additional capacity should be allocated dynamically. This paper proposes a new scheme for dynamic polling that considers both the overhead of polling traffic and the message rates required for management tasks. Our scheme examines polling traffic, applying the discrete Fourier transformation to extract the desired polling rate. We demonstrate the availability of this approach through simulations in which polling requests are varied dynamically. Our scheme adjusts polling rates to make them more consistent and to reduce resource overheads.