Atay Ozgovde

Distributed and Online Fair Resource Management in Video Surveillance Sensor Networks

Visual capability introduced to Wireless Sensor Networks (WSNs) render many novel applications that would otherwise be infeasible. However, unlike legacy WSNs which are commercially deployed in applications, visual sensor networks create additional research problems that delays the real-world implementations. Conveying real-time video streams over resource constrained sensor hardware remains to be a challenging task. As a remedy, we propose a fairness-based approach to enhance the event reporting and detection performance of the Video Surveillance Sensor Networks. Instead of achieving fairness only for flows or for nodes as investigated in the literature, we concentrate on the whole application requirement. Accordingly, our Event-Based Fairness (EBF) scheme aims at fair resource allocation for the application level messaging units called events. We identify the crucial network-wide resources as the in-queue processing turn of the frames and the channel access opportunities of the nodes. We show that fair treatment of events, as opposed to regular flow of frames, results in enhanced performance in terms of the number of frames reported per event and the reporting latency. EBF is a robust mechanism that can be used as a stand-alone or as a complementary method to other possible performance enhancement methods for video sensor networks implemented at other communication layers.

Phone position/placement detection using accelerometer: Impact on activity recognition

Smart phone platforms, equipped with a rich set of sensors enable mobile sensing applications that support users for both personal sensing and large-scale community sensing. In such mobile sensing applications, the position/placement of the phone relative to the user body provides valuable context information. For example, in physical activity recognition using motion sensors, the position of the phone provides important information, since the sensors generate different signals when the phone is carried in different positions and this makes it difficult to successfully identify the activities with sensor data coming from different positions. In this paper, we investigate whether it is possible to successfully identify phone positions using only accelerometer data which is the most commonly used sensor on physical activity recognition studies, rather than using additional sensors. Additionally, we explore how much this position information increases the activity recognition accuracy compared with position independent activity recognition. For this purpose, we collected activity data from 15 participants carrying three phones in different positions, performing activities of walking, running, sitting, standing, climbing up/down stairs, transportation with a bus, making a phone call, interacting with an application on the smart phone, sending an SMS. The collected data is processed with the Random Forest classifier. According to the results of position recognition, using basic accelerometer features which are also used in the activity recognition, can achieve an accuracy of 77.34%, however, this ratio increases to 85% when basic features are combined with angular features calculated from the orientation of the phone. According to the results of the activity recognition experiments, on average the results are similar for position specific and position independent recognition. Only for the pocket case, 2% increase was observed.

WCOT: A utility based lifetime metric for wireless sensor networks

Network lifetime is a novel performance metric which is derived in need to evaluate the networks that are composed of nodes with non-replenishable energy sources. Wireless Sensor Networks (WSNs) are the primary examples of such networks, in which elongating the network lifetime is the main concern. Optimal WSN design is highly dependent on the application scenario context. Correct quantification of the application specific network lifetime is a must to further optimize the design or to comparatively evaluate the proposed schemes - e.g. a legacy layered design vs. a cross-layer implementation. However, in practice, we observe that the focus is given on proposing sophisticated schemes to increase the energy efficiency, whereas only rudimentary lifetime metrics are employed to evaluate the outcome of this effort which compromises the correctness of the results. To realistically and correctly quantify the lifetime, we propose a utility based lifetime measurement framework called Weighted Cumulative Operational Time (WCOT). WCOT lets users incorporate the application dependence into the lifetime metric through its utility based interface. WCOT performs a weighted summation of time where utility values are the weights. With this mechanism, a more representative lifetime metric which maps the complete network behavior into a numeric value is obtained. This is in contrast with metrics which focus solely on certain milestones of the network functionality to quantify the lifetime which include the first node death, the last node death.

Effect of sleep schedule and frame rate on the capabilities of Video Sensor Networks

Visual data carried on video sensor networks (VSNs) creates unique challenges due to the limited hardware of the sensors. In this paper, we study the capabilities of VSNs using currently available technology by investigating how the duty cycle of the sleep schedule and the frame rate affect the overall performance of a VSN. The trade-off between sensor video quality and the received frame rate at the sink is depicted which is an important measure for the application level performance. Using extensive simulations run with realistic parameters, it is shown that increasing the video quality to enhance the application level detection and/or tracking performance works only up to a certain extent.

WCOT: A Realistic Lifetime Metric for the Performance Evaluation of Wireless Sensor Networks

Wireless Sensor Networks (WSNs) give rise to a new networking paradigm in which energy efficiency is a high priority goal. A direct measure of the energy efficiency is the network lifetime which WSN proposals strive to extend. To correctly quantify the lifetime, the metric must be defined in an application dependent manner. In this paper, we propose a generic lifetime measurement framework called Weighted Cumulative Operational Time (WCOT) for the performance evaluation of the WSNs. Novelty brought by WCOT is twofold: First, it defines a utility based interface for the diverse WSN applications to incorporate their scenario specific requirements into the metric itself. Second, WCOT assigns different weights to the operational durations that have different utilities and perform a weighted summation to calculate the cumulative lifetime thereafter. With this mechanism, a more representative lifetime metric which maps the complete network behavior into a numeric value is obtained. This is in contrast with metrics which focus solely on certain milestones of the network functionality to quantify the lifetime which include the first node death, the last node death.

EdgeCloudSim: An environment for performance evaluation of Edge Computing systems

Edge Computing is a fast growing field of research covering a spectrum of technologies such as Cloudlets, Fog Computing and Mobile Edge Computing (MEC). Edge Computing involves technically more sophisticated setup when compared with the pure Cloud Computing and pure Mobile Computing cases since both computational and network resources should be considered simultaneously. In that respect, it provides a larger design space with many parameters rendering a variety of novel approaches feasible. Given the complexity, Edge Computing designs deserve scientific scrutiny for sound assessment of their feasibility. However, despite increasing research activity, this field lacks a simulation tool compatible with the requirements. Starting from available simulators a significant programming effort is required to obtain a simulation tool meeting the actual needs. To decrease the barriers, a new simulator tool called EdgeCloudSim streamlined for Edge Computing scenarios is proposed in this work. EdgeCloudSim builds upon CloudSim to address the specific demands of Edge Computing research and support necessary functionality in terms of computation and networking abilities. To demonstrate the capabilities of EdgeCloudSim an experiment setup based on different edge architectures is simulated and the effect of the computational and networking system parameters on the results are depicted.

Energy Enhancement of Multi-application Monitoring Systems for Smart Buildings

High energy consumption of sensor devices is a major problem in smart building systems, since it strongly impacts the system lifetime. However, existing approaches are often fitted to a single monitoring application and rely on static configurations for sensor devices: optimization of their acquisition and transmission frequencies to actual multiple application requirements is not tackled. In this paper, we focus on energy-aware dynamic sensor device re-configuration to lower energy consumption while fulfilling real-time application requirements. We introduce the Smart-Service Stream-oriented Sensor Management (3SoSM) that binds together sensor configuration and management of sensor data streams. We present a multi-application monitoring system architecture that optimizes application requirements for data streams into sensor device configurations, and we relate the experiments with our experimental platform.

Enabling service-centric networks for cloudlets using SDN

The current developments in smart devices, wearable gadgets and IoT (Internet-of-Things) are triggering a variety of novel use cases and services. Once the technology matures, a wide range of services is expected to be provided at the edge of the network in coordination with the cloud computing infrastructure. These services will be highly dynamic meaning that they can be served from edge computing facilities and from central cloud servers being transferred back and forth. Also, considering the mobility of the users and the varying demand, those services might need to be live migrated between nearby edge servers on-the-fly. This setup creates an environment that needs to be transparent to the end users. The current legacy networking paradigm, however, allow services to be reached by IP and port addresses, not by their content. Alternatively, in the service-centric approach the services themselves are handled independent of their location and the focus shifts to “what ” instead of “where ”. Due to the complexities involved, it is not a straightforward task to establish service-centricity at the edge scenarios. As a remedy, this paper proposes a service-centric approach at the edge servers using orchestration capabilities offered by the Software-Defined Networking (SDN) technology. To demonstrate how SDN can help to alleviate the problem, an emulation environment is used in which northbound applications are implemented in order to setup the service-centric structure. The effect of service-centric approach is shown with the load balancing experiments where the performance of the proposed system is evaluated for various use cases.

Application specific dynamic sleep scheduling

Ambient Intelligence is one of the research area where wireless sensor devices are commonly used. Main idea of these applications is to monitor and recognize peoples indoor or outdoor activities. In this study, we focus on one of the major problem of WSN which is energy consumption in a different way from the existing approaches in the literature. With the proposed approach, it is shown that it is possible to create a dynamic sleep scheduling mechanism for each recognized activity. In contrast to previous studies found in the literature review, in our perspective, a holistic approach which takes component and application level into account is proposed, therefore energy consumption and the ways for improving lifetime of node and the network are handled. Main aim is to find a suitable region where application can work with tolerable performance degradation meanwhile lifetime of the sensor is extended by decreasing energy consumption. Proposed approach is tested on accelerometer signal of walking activity. It is shown that it exists a region where application performance and energy gain are suitable.

Position-aware activity recognition on mobile phones

Activity recognition (AR) or in other words context recognition is an active area of research in the domain of pervasive and mobile computing that has direct applications about life quality and health of the users. Related studies aim to classify different daily human activities with high accuracy rates using various types of sensors. Becoming a substantial part in our daily lives with their sensing capabilities, smartphones are now feasible platforms that enable people to make use of AR technologies without being obliged to use or wear some extra devices. However, due to the fact that users carry these devices at different positions, such as in the pocket or in the bag, it becomes a challenging task to attain accurate results by directly used classification models. In this paper, we focus on phone position uncertainty problem and compare the classification results with position independent and position dependent classification models.