Collecting and processing data from wireless sensor networks

Collecting and processing data from wireless sensor networks Pavel Vajsar 1, Patrik Morávek 1, Petr Žoldoš 1 1 Faculty of Electrical Engineering and Communication Brno University of Technology Email: pavel.vajsar@phd.feec.vutbr.cz, moravek@phd.feec.vutbr.cz, xzoldo00@stud.feec.vutbr.cz Abstract This article describes solution called WSN 2 Monitor for monitoring wireless sensor networks. The basic of the WSN Monitor solution is database storage. Design of database storage is based on analyses of real wireless sensor network parameters, which are important for monitoring. Application is the main part of the solution. It is divided into client-side application and server-side application. The application is flexible and scalable for new module features. The client-side application is fully supported by mobile devices. 1 Introduction This paper is organized into six chapters. Chapter one introduces the basic goal of the project. Technologies chosen for development and software architecture are described afterwards. Third capture explains the way the parts of the system structure work and cooperate with each other. Graphical printscreen and diagrams are included for better understanding as well. Chapter number four focuses on handheld devices support and at the end the conclusion of the final product is made. Wireless sensor network (WSN) technology has recently gained a great interest of scientists and researchers [1], which is still increasing. New ways of deployment, routing, data gathering and applications are being developed for this new technology. Critical part of each wireless sensor network is graphical user tool allowing effective management, control and evaluation, which is used by user and operator. It takes important place in each control center. Because of variety of types and vendors of WSN products it was necessary to evaluate and collect information, which is common to most of them. This information formes the basis of the storage design, which is capable of collecting all product related data. Storage design is not sufficient enough to monitor and evaluate network states effectively, therefore the emphasis was placed on the application design, which is capable of working with pre-designed storage much more effectively. One of many requirements on the solution was to support the majority of highly used operating system platforms (e.g. MS Windows, GNU Linux, Mac OS) and also some of mobile operating system platforms (e.g. Android). Monitoring application was designed as easily extended featuring modular architecture and as consequence of rapid WSN technology development and wide range of WSN applications [2-6]. Most of the producers of wireless sensor network components provide software, which is able to manage and monitor 2 WSN is a shortcut for Wireless Sensor Network wireless sensor network. This software is usually useful only for manufacturer s products. Usage is often limited by the type of operating system platform. In most cases, the MS Windows platform is supported because it is the most common platform. Those kinds of monitoring software are for example imonnit TM Online Sensor Monitoring [7] or Mote-View [8]. It is very complicated and sometimes even impossible to extend the software by new features. Mobile devices are usually not supported at all. The solution called WSN Monitor is not linked to any particular network, it was being tested in experimental environment of real wireless sensor network. [9] 2 Solution Pre-requirements WSN Monitor consists of database storage, server-side application and client-side application. At the time the solution was being designed, it was really important for each part of it to support the most used operating system platforms. Furthermore, minimal requirements of user s software and hardware equipment, broad availability, high stability, low-cost administration and maintenance. The software tools used to develop each part of the solution were selected with regard to our demands and the price of the license. Free tools were preferred. MySQL database storage was chosen for storing data from wireless sensor network. Apache Tomcat was chosen as runtime environment for server-side application. Server side application was built on Java technology with support of several frameworks such as Hibernate [10], Spring [11], Spring Security [12], BlazeDS [13] and others. Client side application is based on Adobe Flex [14] / Adobe Air technology. [15] 3 Application architecture The database storage was designed to store the information needed for network control and evaluation of network state. Flexibility of the database was considered as the main requirement. The proposal was proceeded by analysis of parameters from various wireless sensor networks. The database can store monitor and maintain values such as: Complete information about each node in network. Information about node neighbours, presence in a cluster and size of the cluster. Information about wireless radio link with neighboring nodes. 1

Description of environment in which the wireless sensor network has been deployed. Detail information about position of each node in network (building, area, room). In addition the database stores information, which is not related directly to wireless sensor network, but is important for the functionality of the application such as user accounts, roles, login/logout information. WSN Monitor using proposed database repository is divided into two main parts: Server-side application Client-side application The WSN Monitor is divided into two parts. Each of parts has different requirements on runtime environment. The server side application requires environment with Java Servlet and JavaServer Pages technologies support. The solution was during developing using Apache Tomcat server, which supported required technologies. The Apache Tomcat server is released under the Apache License version two which is compatible with the GPL. Frameworks support in the server side application is achieved by copy framework`s libraries into application build path. The client side application based on Adobe Flex technology requires Adobe Flash player installed into client`s web browser. The client-side application based on Adobe Air technology requires Adobe Air runtime installed on operating system. The entire structure is shown in the Figure 1. The serverside application accesses data and processes requests from client-side application. Server-side application can read, write and modify data in database storage. These operations are implemented through services. Access to database is executed, when client side application calls some of these services. Server side application also provides security. Services use Hibernate framework, which provides object-relational mapping. Object-relational mapping is a technique for data conversion between server-side application and database. This technique offers independent relation between application and database storage. The only information needed consequently is information regarding database server address and type of database server. It means that the application can use any type of database storage, which is supported by Hibernate framework. Designed database structure is written into Java objects. If the application is executed for the first time and the database structure does not exist yet, the database object model is used to create the relational database structure. Server-side application also provides security. This feature is provided by Spring Security framework. The application verifies the username and password. Username and password are both related to an user account which is stored in database. Hash function SHA-256 is used to save the password into the database and to verify user s identity. Client-side application provides user interface containing graphical components and control items. Client and serverside applications are separated and independent. Because of that, it is possible for each application to run on different server. Moreover the client-side application can be built on another technology instead of Adobe Flex / Adobe Air as well, what brings even higher portability. Communication between both applications is ensured by BlazeDS. 3.1 Application requirements Figure 1. Application Architecture WSN Monitor has modular structure. Each module of application provides particular functionality of network monitoring. Modules are mutually independent and are loaded into memory when necessary. If a module is not currently used it is unloaded. Unloading means that the module is removed from the memory and system resources are released. Therefore, only one module takes memory space at each particular time the module, which is currently being used. If any error occurres in any module, it will not cause an error in the whole application, but only in the specific module. Other functions and other modules will not be influenced and will work as they suppose to. During the solution development the emphasis was placed on the possibility of extension and customization which was achieved by dynamically loaded modules. User interface is designed with the emphasis on the ease of use and is divided into three main parts: 2

Figure 2. Application user interface (Real Time module) Top panel contains information about user logged in the application, further it contains button for full screen mode and button for log off from application. Module panel this panel displays actually loaded module. The panel title contains path and name of actually loaded module. Icon panel the panel is located at the bottom part of the screen and contains icons of each available module. data from database to CSV or PDF format. Exported data can be used for further detailed analyses in other software tools, such as MathWorks or Matlab, can be printed. Figure 2 shows application user interfaces. Addition real time module can be seen in module panel. Real time module was developed specifically for real time monitoring of sensor network. It is able to connect to a certain sensor network and receive data from the network. Thereafter, the data are analyzed, divided into basic parts and stored in the database repository. Connection with real sensor network is realized through Xserve server. Xserve provides wireless sensor network gateway. The base station, which collects data from the network, is connected to the Xserve server. The real time module is connected to Xserve. Communication structure is shown in Figure 3. WSN Monitor has extra modules, which provide information about geographical network structure (placement of the node in a deployment area), sensors and last module analyzes data from wireless sensor network and exports measurement Figure 3. Communication structure Map Module - The Map module shows position of all wireless sensor nodes. Google maps SDK, which is used to display each node on the map, works as map support. The module provides basic information about each node shown on the map. After mouse movement over the node the information table is revealed. 3

Graphic Module - Graphic module is used to display data from wireless sensor network, which are stored in database storage graphically. Data can be exported to CSV file and charted to PDF file. Sensor Module Sensor module provides information about each node of wireless sensor network. Related information are grouped into folders for convenience when using the module. Most important information includes available sensor devices and measurement categories. Module also provides information about hardware state or power for example. User Module - User Module operates with basic user related information and allows currently logged user to change his or her own personal information and password to access the application. It also provides information about last login into the application. 4 Mobile devices support Wireless sensor network can be deployed in diverse conditions and environments where it may be necessary to monitor the network right on site. Particularly for these purposes Adobe Air technology support was introduced. This technology is supported in the majority of mobile devices. The client-side application is uploaded into a mobile device. It is loaded from mobile device after application startup, not from Internet as Adobe Flex version running in web browser. The advantage is no need of downloading the application from the Internet. This solution increases the speed of application lunching in mobile devices and saves its energy. On the other hand manual version update is considered as disadvantage of this solution. Communication structure is shown in Figure 4. Figure 4. Mobile access 5 Conclusion This article describes the solution proposed for monitoring wireless sensor networks. The important feature of this solution is the support of most of commonly used operating system platforms, as well as the support of some mobile operating system platforms. The solution is scalable because of its modular architecture deploying modules. Modules can be created and modified based on specific needs and requirements. The application has already been tested during the experimental measurements and it successfully handled high data load. Application is able to capture packets from network, decode data, store captured data into the database and present graphical output of monitored value. The values obtained from the wireless sensor network can be exported to CSV format, which can be opened in spreadsheet applications. Charts can be exported into PDF format. Further work consists of improvement of server side application with modules support. The work will be focused on improve. Currently the module is showing detailed network model. Module will provide information about network topology inside the building as well as outdoor. The XML structure was designed as data provider for structure, which can be imported or exported to / from application. 6 References [1] MAINWARING, Alan. Wireless sensor networks for habitat monitoring. WSNA '02 Proceedings of the 1 st ACM international workshop on Wireless sensor networks and applications. 2002. [2] MORE, Darren. Monitoring Rainforest Regeneration. CSIRO 2009, http://research.ict.csiro.au. [3] Dr DUNBABIN, Matt. Water Quality Monitoring. CSIRO 2009, http://research.ict.csiro.au. [4] SWARTZ, R. Andrew. Structural Monitoring of Wind Turbines using Wireless. Cambridge. [5] XU, Ning, A wireless sensor network For structural monitoring. In SenSys '04 Proceedings of the 2 nd international conference on Embedded netwo r- ked sensor systems, 2004. [6] Werner-Allen, G.; Johnson, J.; Ruiz, M.; Lees, J.; Welsh, M.. Monitoring volcanic eruptions with a wireless sensor network. Proceeedings of the Second European Workshop. 2005. [7] Monnit. 2011. Wireless Sensor Network Monitoring and Notification Software. 2011 http://www.monnit.com/products/monitoringoptions.php. [8] WILLOW Technologies. Mote-View Client Software. 2006. http://www.willow.co.uk/html/moteview_client_software.html. [9] Limin Sun; Yuyan Sun; Jian Shu; Qingwei He. MotePlat: A Monitoring and Control Platform for Wireless Sensor Networks Grid and Cooperative Computing Workshops, GCCW Fifth International Conference Oct. 2006. 4

[10] JBoss Community. Hibernate. 2010. http://www.hibernate.org/ [11] Spring Source. Spring Framework. 2010. http://www.springsource.com [12] Spring Source. Spring Security. 2010. http://static.springsource.org/spring-security/site/ index.html. [13] Adobe Systems Incorporated. BlazeDS. Adobe- Open Source. 2008. http://opensource.adobe.com/ wiki/display/blazeds/blazeds/. [14] Adobe Systems Incorporated. Adobe FLEX Technology. 2010. http://labs.adobe.com/products/flex/. [15] Adobe Systems Incorporated. Adobe AIR Technologies. 2010. http://www.adobe.com/technologies/air/. 5