Michal Marks

Optimization Schemes For Wireless Sensor Network Localization

Optimization Schemes For Wireless Sensor Network Localization Many applications of wireless sensor networks (WSN) require information about the geographical location of each sensor node. Self-organization and localization capabilities are one of the most important requirements in sensor networks. This paper provides an overview of centralized distance-based algorithms for estimating the positions of nodes in a sensor network. We discuss and compare three approaches: semidefinite programming, simulated annealing and two-phase stochastic optimization—a hybrid scheme that we have proposed. We analyze the properties of all listed methods and report the results of numerical tests. Particular attention is paid to our technique—the two-phase method—that uses a combination of trilateration, and stochastic optimization for performing sensor localization. We describe its performance in the case of centralized and distributed implementations.

HETEROGENEOUS GPU&CPU CLUSTER FOR HIGH PERFORMANCE COMPUTING IN CRYPTOGRAPHY

This paper addresses issues associated with distributed computing systems and the application of mixed GPU&CPU technology to data encryption and decryption algorithms. We describe a heterogenous cluster HGCC formed by two types of nodes: Intel processor with NVIDIA graphics processing unit and AMD processor with AMD graphics processing unit (formerly ATI), and a novel software framework that hides the heterogeneity of our cluster and provides tools for solving complex scientific and engineering problems. Finally, we present the results of numerical experiments. The considered case study is concerned with parallel implementations of selected cryptanalysis algorithms. The main goal of the paper is to show the wide applicability of the GPU&CPU technology to large scale computation and data processing.

High performance wireless sensor network localisation system

In this paper we summarise the results of our research concerned with the development, implementation and evaluation of a software framework for wireless sensor networks (WSNs) localisation - high performance localisation system (HPLS). The system can be used to calculate positions of sensing devices (network nodes) in the deployment area, and to tune and verify various localisation schemes through simulation. It provides tools for data acquisition from a workspace, estimation of inter-node distances, calculation of geographical coordinates of all nodes with unknown position and results evaluation. Received Signal Strength measurements are utilised to support the localisation process. Trilateration, simulated annealing (SA) and genetic algorithm (GA) are applied to calculate the geographical coordinates of network nodes. The utility, efficiency and scalability of the proposed localisation system HPLS have been justified through simulation and testbed implementation. The calculations have been done in parallel using the map-reduce paradigm and the high performance computing (HPC) environment formed by a cluster of servers. The testbed networks were formed by sensor devices manufactured by Advantic Technology (clones of TelosB platform). A provided case study demonstrates the localisation accuracy obtained for small-, medium and large-size multihop networks.

Hybrid CPU/GPU Platform For High Performance Computing.

High performance computing is required in a number of data-intensive domains. CPU and GPU clusters are one of the most progressive branches in a field of parallel computing and data processing nowadays. Cloud computing has recently emerged as one of the buzzwords in the ICT industry. It offers suitable abstractions to manage the complexity of large data processing and analysis in various domains. This paper addresses issues associated with distributed computational system and the application of mixed GPU&CPU technology to data intensive computation. We describe a hybrid cluster formed by devices from different vendors (Intel, AMD, NVIDIA). Two variants of software environment that hides the heterogeneity of our hardware platform and provides tools for solving complex scientific and engineering problems are presented and discussed. The first solution (HGCC) is a software platform for data processing in heterogenous CPU/GPU clusters. The second solution (HGCVC) is an extension version of the previous one. The cloud technology is incorporated to the HGCC framework. The results of numerical experiments performed for parallel implementations of password recovery algorithms are presented to illustrate the performance of our systems.

A Movement-Assisted Deployment of Collaborating Autonomous Sensors for Indoor and Outdoor Environment Monitoring

Using mobile robots or unmanned vehicles to assist optimal wireless sensors deployment in a working space can significantly enhance the capability to investigate unknown environments. This paper addresses the issues of the application of numerical optimization and computer simulation techniques to on-line calculation of a wireless sensor network topology for monitoring and tracking purposes. We focus on the design of a self-organizing and collaborative mobile network that enables a continuous data transmission to the data sink (base station) and automatically adapts its behavior to changes in the environment to achieve a common goal. The pre-defined and self-configuring approaches to the mobile-based deployment of sensors are compared and discussed. A family of novel algorithms for the optimal placement of mobile wireless devices for permanent monitoring of indoor and outdoor dynamic environments is described. They employ a network connectivity-maintaining mobility model utilizing the concept of the virtual potential function for calculating the motion trajectories of platforms carrying sensors. Their quality and utility have been justified through simulation experiments and are discussed in the final part of the paper.

Software environment for parallel optimization of complex systems

The paper is concerned with parallel global optimization techniques that can be applied to solve complex optimization problems, and are widely used in applied science and in engineering. We describe an integrated software platform EPOCS (Environment for Parallel Optimization of Complex Systems) that provides the framework and tools which allow to solve complex optimization problems on parallel and multi-core computers. The composition, design and usage of EPOCS is discussed. Next, we evaluate the performance of methods implemented in the EPOCS library based on numerical results for a commonly used set of functions from the literature. The case study --- calculating the optimal prices of products that are sold in the market is presented to illustrate the application of our tool to a given real-life problem.

Localization based on stochastic optimization and RSSI measurements

The paper describes the design and performance of a novel technique that can be used to calculate the geographic locations of nodes that form wireless sensor network system (WSN). Location awareness is required for many WSN applications, but it is often too expensive to include GPS adapter in each sensor node. Hence, localization systems are provided to calculate the positions of nodes. We developed multihop algorithms based on trilateration and stochastic optimization, and RSSI measurements.

Network-wide power management in computer networks

An important part of the modern computer networks design is to develop novel technologies, architectures and control mechanisms for network devices enabling power saving by adapting network capacities to current traffic loads and user demands. We describe centralize and hierarchical control frameworks for reducing power consumption in backbone computer networks. The implementation of these frameworks provides the local control mechanisms that are implemented in the network devices level and network-wide control strategies implemented in the central control level. In this paper, we focus on network-wide algorithms for calculating the power status of network devices and the energy-aware MPLS routing for recommended network configuration. We enumerate several possible formulations of a network energy saving optimization problem with continuous and discrete variables. We discuss the limitations of these approaches and problems with their application to power control in real networks. We propose the relaxation of the complete binary problem formulation assuming full routing and energy state of all devices calculation, and the algorithm to solve it. Our formulation is based on a heuristic approach that leads to a continuous optimization. The evaluation of the optimization scheme through simulation is presented in the final part of the paper.

Real Life Data Acquisition In Wireless Sensor Network Localization System.

The paper treats the problem of localization in Wireless Sensor Network (WSN). In our work, we present and evaluate Wireless Sensor Network Localization System, which supports sensor node localization from data gathering from real-life deployments through modelling and applying different localization methods up to distributed computing in HPC environment. The paper describes extension of WSN Localization System with modules supporting real-life sensor data acquisition. A provided case study demonstrates the localization accuracy obtained for a few example networks generated by simulation models and based on acquired sensor data. INTRODUCTION TO WSN LOCALIZATION The aim of localization is to assign geographic coordinates to each node in the sensor network in the deployment area. Wireless sensor network localization is a complex problem that can be solved in different ways, [Karl and Willig, 2005]. A number of research and commercial location systems for WSNs have been developed. They differ in their assumptions about the network configuration, distribution of calculation processes, mobility and finally the hardware’s capabilities, [Mao et al., 2007], [Awad et al., 2007], [Zhang et al., 2010]. Recently proposed localization techniques consist in identification of approximate location of nodes based on merely partial information on the location of the set of nodes in a sensor network. An anchor is defined as a node that is aware of its own location, either through GPS or manual pre-programming during deployment. Identification of the location of other nodes is up to an algorithm locating non-anchors. Considering hardware’s capabilities of network nodes we can distinguish two classes of methods: range based (distance-based) methods and range free (connectivity based) methods. The former is defined by protocols that use absolute point to point distance estimates (ranges) or angle estimates in location calculation. The latter makes no assumption about the availability or validity of such information, and use only connectivity information to locate the entire sensor network. The popular range free solutions are hop-counting techniques. Distancebased methods require the additional equipment but through that much better resolution can be reached than in case of connectivity based ones. In our works we concentrate on range based methods. The paper is structured as follows: at the beginning we shortly describe the distance-based localization problem. Next, we provide an overview of our software environment for WSN localization and an extension applied to our software in order to acquire data from real-life deployments. Finally, we provide a case study results and conclusions. DISTANCE BASED LOCALIZATION Let us consider a network formed by M sensor devices (anchor nodes) that are aware of their location, either through GPS or manual recording and entering position during deployment, and N sensor devices (nonanchor nodes) that are not aware of their location in a network system. The goal of a localization system is to estimate coordinate vectors of all N non-anchor nodes. In general, distance based localization schemes operate in two stages: • Inter-node distances estimation stage – estimation of true inter-node distances based on inter-node transmissions and measurements. • Position calculation stage – transformation of calculated distances into geographic coordinates of nodes forming the network. Inter-node Distances Estimation Stage In spite of the available hardware, distance based localization systems exploit the following techniques Proceedings 27th European Conference on Modelling and Simulation ©ECMS Webjørn Rekdalsbakken, Robin T. Bye, Houxiang Zhang (Editors) ISBN: 978-0-9564944-6-7 / ISBN: 978-0-9564944-7-4 (CD) Fig. 1. The components of the WSNLS widely described in literature [Benkic et al., 2008], [Karl and Willig, 2005], [Mao et al., 2007]: • Angle of Arrival (AoA), • Time of Arrival (ToA), • Time Difference of Arrival (TDoA), • Received Signal Strength Indicator (RSSI). AoA, ToA and TDoA methods need an additional equipment such as antennas or accurately synchronized clocks. The most popular technique is the RSSI method because of easy configuration, deployment and no additional hardware needed (low cost). The disadvantage of this solution is low quality of measurement accuracy due to high variability of RSSI value [Benkic et al., 2008], [Ramadurai and Sichitiu, 2003]. Nevertheless some authors indicate that new radio transceivers can give RSSI measurements good enough to be a reasonable distance estimator [Barsocchi et al., 2009], [Srinivasan and Levis, 2006]. Position Calculation Stage In the position calculation stage the computed internode distances are used to estimate the geographic coordinates of all non-anchor nodes in a considered network. Position estimation can be done by using different techniques. There are many widely used techniques such as: triangulation, trilateration, multitrilateration and multidimensional scaling. The common idea of other methods is formulating the localization problem as the linear, quadratic or nonconvex nonlinear optimization problem solved by linear, quadratic or nonlinear (often heuristic) solvers. Recently, a popular group consists of hybrid systems that combines more than one technique to estimate location, i.e., results of initial localization are refined using another localization method. All mentioned methods are described and evaluated in literature, see [Akyildiz and Vuran, 2010], [Biswas and Ye, 2004], [Kannan et al., 2005], [Kannan et al., 2006], [Mao and Fidan, 2009], [Mao et al., 2007], [Niewiadomska-Szynkiewicz, 2012], [Niewiadomska-Szynkiewicz et al., 2011]. WIRELESS SENSOR NETWORK LOCALIZATION SYSTEM OVERVIEW The Wireless Sensor Network Localization System (WSNLS) is an integrated software environment for testing various localization schemes on parallel computers or computer clusters. It provides not only a set of solvers for localization WSN nodes but supports the whole localization process from test network defining, radio signal modelling and processing, real-life data acquisition up to parallel execution of localization schemes. An open architecture and object-oriented programming make the software easily extendable with implementations of new approaches for calculating locations of nodes in a network. WSNLS can be used to estimate the geographic coordinates of all devices forming the real life sensor network. Moreover, it can be used for tuning and performance evaluation of various localization solvers that are integrated with the framework before their practical application to a real life network. Since its first realization, described in [Marks, 2012], WSNLS architecture has been improved in many aspects and extended by adding Sensor Data Acquisition Module (SDAM). The system is still composed of a runtime platform (formed by two components, i.e., Distributed Computing Manager and Computational Server) responsible for calculation management and interprocess communication. However in second version of the system, Networks Manager has been reorganized and contains Networks Generator – a component for modeling a network to be simulated and Sensor Data Fig. 2. Dataflow in WSNL system Acquisition Module – component responsible for data gathering from real-life deployments. There are still two components responsible for location calculations, i.e., Distance Estimation Module and Position Calculation Module, database for recording data of all examined networks and results of calculations, and a set of tools to support the interaction with a user (GUI), but all the features of Position Calculation Module are realized by computational servers. The architecture of WSNLS is presented in Fig. 1. DATA FLOW IN WSNLS Since the aim of WSNLS is providing support for the whole localization process – from test network defining up to nodes location estimation – the data processing requires applying specialized methods on three stages as it is shown in Fig. 2. Computational method used on two stages i.e. distance estimation methods and position calculation methods are described in more details in [Marks, 2012], [Marks and NiewiadomskaSzynkiewicz, 2011]. However the first stage in presented dataflow relies on topology estimation methods, which were partially unavailable in first version of Wireless Sensor Network Localization System. Topology estimation methods Topology estimation methods provide a means for gathering information about network topology. This information can be obtained by using appropriate modelling or by data acquisition from real-life deployments. In general the proper modeling of low-power links is very difficult since the links characterization depends on radio chips (e.g., TR1000, CC1000, CC2420, etc), operational environments (indoor, outdoor) and many other parameters such as traffic load or radio channel – [Baccour et al., 2012]. In our software we decided to provide models based on Link Layer Model for MATLAB provided by [Zuniga and Krishnamachari, 2004], where we focus on wireless channel modeling and no radio modulation and encoding are considered. Much better solution, of course applicable only for institution which have at least laboratory WSN deployments, is to acquire data directly from real Wireless Sensor Networks. More information about real-life data acquisition is provided in section Sensor Data Acquisition Module. Distance estimation methods Distance estimation methods transform RSSI measurements into internode distances estimations. At present Distance Estimation Module has registered three approaches to distance estimation: Ordinary Least Square Method (OLS), Weighted Least Square Method (WLS) and Geometric Combined Least Square Method (GCLS). More information about distance estimation stage can be found in [Marks and Niewiadomska-Szynkiewicz, 2011]. Position calculation methods Position calculation methods estimate the coordi

Poster Abstract: WSNLOC.EU - An Introduction to WSN Localization Tasks Repository

The paper introduces the WSN Localization Task Repository - wsnloc.eu, service structure and our motivations why it was created.