Melike Erol

AUV-Aided Localization for Underwater Sensor Networks

We propose a localization scheme for underwater acoustic sensor networks (UWSN) that does not require a priori infra-structure or synchronization between nodes. An autonomous underwater vehicle (AUV) aids in localizing the sensor nodes while roaming across the underwater sensor field. The objectives of this paper are to describe how to localize nodes using AUV and to describe the tradeoffs involved, i.e. ratio of localized nodes and localization accuracy. We show that localization success improves as the duration of the AUV localization process increases. In addition, we investigated localization using two methods, bounding-box and triangulation. The former achieves a higher localization ratio but with a higher error. In certain scenarios, we achieved 100% nodes localized with 3% error.

The Meandering Current Mobility Model and its Impact on Underwater Mobile Sensor Networks

Underwater mobile acoustic sensor networks are promising tools for the exploration of the oceans. These networks require new robust solutions for fundamental issues such as: localization service for data tagging and networking protocols for communication. All these tasks are closely related with connectivity, coverage and deployment of the network. A realistic mobility model that can capture the physical movement of the sensor nodes with ocean currents gives better understanding on the above problems. In this paper, we propose a novel physically-inspired mobility model which is representative of underwater environments. We study how the model affects a range-based localization protocol, and its impact on the coverage and connectivity of the network under different deployment scenarios.

Localization in Wireless Sensor Networks with Range Measurement Errors

In sensor networks, data collected by sensor nodes needs to be tagged with time and location information. Localization techniques are used to determine the location information by estimating location of a sensor node. It is well known that distance measurement errors affect the accuracy of estimated location. These errors may be due to methodical or environmental factors. In this paper, we propose AML (Adapted Multi-Lateration) by improving the existing multi-lateration technique. It is shown that the AML method is more robust to measurement errors; its mean localization error is lower than the multi-lateration technique for noisy measurements. Besides, the time complexity of the AML method is less than the multi-lateration technique since it does not require to solve the normal equation for the linear least squares problem as in the multi-lateration technique. Additionally, AML is advantageous for iterative localization where localized nodes become reference nodes and employed in the localization process. Incorporating these reference nodes in the AML equations is easier than multi-lateration technique.

Analysis of multi-player online game traffic based on self-similarity

When provisioning network resources, ISPs have to tackle with the difficulties imposed by the traffic characteristics and demands of the users. Multi-player online games (MPOG) are constituting a large portion of the Internet traffic. MPOG players demand high bandwidth and low delay due to high interaction in these games. Moreover, they generate bursty traffic. In this paper, we analyze the traffic characteristics of a popular MPOG type game; Call of Duty 2 version 1.0 (CoD2). We use several game sessions to extract packet size and inter-arrival time distributions. In addition to these we investigate the self-similarity of the CoD2 traffic. We observe two different behavior at server and client-side, as expected. For the server-side, traffic characteristics vary with the number of players, but the traffic is long-range dependent most of the time. There are several intervals when the traffic process is non-stationary where the analysis of self-similarity becomes trivial. For the client-side, the traffic shows little variations however there are still some cases where the analysis of self-similarity becomes intricate because the traffic process reveals deviation from stationarity.

Localization in underwater sensor networks

Applications of Underwater Sensors Networks (USN) include environmental ocean/sea monitoring, underwater mine searching, detection of chemically/biologically harmful substances or pollutants, autonomous underwater attack/defense systems, collecting ambient data for ship navigation, etc. The success of terrestrial sensor networks has promising results for more challenging environments, such as underwater. Using groups of sensors communicating with each other yields better performance in underwater applications rather than single sensor equipments. However, forming a network in underwater is not straightforward. The main challenges arise from the physical communication medium and affect the upper protocol stack such as MAC, routing, transport layer protocols and localization. For USNs, protocols at all those layers should be designed with communication and energy cost in mind. Localization is required for data tagging. In terrestrial sensor networks, either GPS is used wherever available or several GPS-free, messaging-intensive schemes have been used. However, both are unsuitable for USNs. In this paper, we compare the performance of two distributed localization methods tailored for large-scale USNs.

A comparative study on the estimation of hurst parameter of fractional Brownian motion

Most natural processes show self-similarity. Self-similarity or 1/f behavior has been a popular concept to describe the scale invariance of time-series. Fractional Brownian motion (fBm) is one of the most preferred models of 1/f processes since statistical behavior of fBm is determined by a single (Hurst) parameter, H. Consequently, estimation of H has been an important issue and several methods are developed to estimate H. It has been a complicated issue to determine which one of the estimators yields more robust estimates. In this work, we compare the performance of three methods, namely, Higuchi, Wavelet Based and recently proposed Principal Component Analysis methods. We apply these estimators to data sets with varying H values. We also analyze the effect of data length on the robustness of the estimators. Finally, we investigate the effects of periodicity interfering with fBm which is a common case encountered in practice. 1. Giriş İstatistiksel olarak öz-benzeşimli bir x(t) rasgele sürecinin zaman ortamındaki istatistiksel özellikleri öteleme ile değişirken, ölçekleme (daraltma-genişletme) ile değişmemektedir. Bu tür süreçler, “” istatistiksel eşdeğerliğin gösterimi, c ölçekleme parametresi olmak üzere, aşağıdaki ölçekleme özelliğine sahiptirler [1]: 0 c , ) ct ( x c ) t ( x H p