- Email: [email protected]
Wireless Sensor Networks and RFID for Improving Information Visibility in Airport Ground Handling Management
Wireless Sensor Networks and RFID for Improving Information Visibility in Airport Ground Handling Management Pablo Garcia Ansola1, Javier de las Morenas1, Andrés García1, Javier Otamendi2, M. de la Fuente Ruz3, Javier Garcia Escribano1 1
Escuela Técnica Superior de Ingenieros Industriales, Universidad de Castilla-La Mancha, Ciudad Real, 13071, Spain. ([email protected]). 2 Universidad Rey Juan Carlos, Facultad de Ciencias Jurídicas y Sociales, Paseo Artilleros, Madrid, 28032, Spain. 3 Departamento de Electrónica, Universidad de Jaén, EPS, Campus las Lagunillas, Jaén, 23071, Spain.
Abstract: The new economic situation in all industries, and especially in the air transportation industry, claims for new business models supported by accurate management processes, which need constant feedback of the real status of the environment. A combined use of identification technologies and wireless sensor networks (WSN) can provide the required visibility to that information on status. This work proposes a WSN based on a Zigbee network of RFID readers with active tags as end nodes. The WSN can get the identification (ID) information stored in the end nodes, which are carried by the physical resources. Besides the ID, the tag can include other useful data coming from sensors such as acceleration, temperature, humidity, fuel level, etc. The final objective is to achieve an updated/real picture of the resources in the airport environment that will help in the decision-making. Keywords: Computer Communication Networks, RFID, Intelligent Manufacturing System.
1. INTRODUCTION New management systems are steadily increasing their complexity due to pressing requirements for productivity and flexibility. This complexity becomes especially poignant when dealing with typically big plants, airports or logistic centers; where the usual control is too rigid and hard to manage, maintain and update (Thorne A., 2007). For a modern system however, finding an adequate response to a broad range of possible new situations must be business-asusual; and this should apply even while the system is being modified/updated (Garcia A., 2003). Thus, distributed control becomes an standard practice as it allows a division of the problem in smaller and more affordable ones. In this case the system becomes distributed; where a negotiation process is relied upon, to take decisions that affect more than one of these parts (E.H., 1996). The need for an improvement in the connection between the real environment and the information systems is essential (T.W., 1999); ERPs (Enterprise Resource Planning) or any other information/management system has given place to new architectures, which need constant feed back of information of the real environment. This information is managed by new artificial intelligent techniques like the Bayesian networks (Jensen, 1996), Multiagent systems (Bussman S., 2004), Neural networks (Grossberg, 1988) or Fuzzy systems (Klir & Yuan, 1995); making expert system adapted to the environment (Russell & Norvig, 1995). To be able to get information from the environment, the systems need a transparent identification technology and quick communications to the information system. RFID technology is perfect to capture information from the physical world; offering rapid responses and continuous tracking. RFID
stands for radio frequency identification; a technology in which there are standards already in use defining communication protocols, hardware and information systems. An example of these standards is the one provided by EPC (Electronic Product Code). However, the reading range of RFID systems is relatively short for some applications, typically from two to three meters for passive tags in UHF and about 8-12 meters for active tags. As known, there are no standards available for a topology composed by several readers as the one typically used for WSN (Akyildiz, Su, Sankarasubramaniam, & Cayirci, 2002). The WSN abstracts the mobility, the topology, the scalability and the communication to the information system (Sohrabi, Gao, Ailawadhi, & Pottie, 2000); besides, the routing information can be added to the information package submitted (Akkaya & Younis, 2005). From this point of view, identification is another way of sensing. The ID is attached to the information of the sensor, getting a complete information package made up of the variables of sensors, the routing path and the unique ID of the sensorized element; making a complete data model in the WSN (Heinzelman, Kulik, & Balakrishnan, 1999). One of the main requirements in the new management is the needed necessity of a dynamic visibility of the environment. The staff does not know where they will need to improve the visibility. It means that the proposed WSN needs to be dynamic in the topology of the network and its information submitted. The second main problem is the needed of a lowlevel consumption in the WSN (Ghiasi, Srivastava, Yang, & Sarrafzadeh, 2002) and the detection (Chen, Kher, & Somani, 2006)& auto diagnostic system (Ssu, Chou, Jiau, & Hu, 2006). The paper will focus on these problems in Chapter 2.
The proposed WSN will be implemented in the Ciudad Real Central Airport in Spain. This is a small airport and its demand is based on low cost companies and cargo operations. The airport wants to improve the productivity of the resources of ground handling trying to reduce delays, overloads and overstaff. That will allow for a reduction of the costs, following the path set by the low cost companies. 1.1 Current situation in airport operations The world economic crisis has forced airlines to modify the way in which they are operating: improving productivity and reducing costs. Modern airports can be extremely complex organizations, which often involve governmental organizations, private companies, airlines, aircraft operations and airport operators. New ideas like common use are exploiting to reduce cost and improve the productivity. The concept of “Common Use Model” indicates that airport operators can gain centralized control over facilities and technologies; increase passenger-processing options; and acquire shared use efficiencies as they move from exclusive use toward common use. Airport common usable space is defined as the space in which any airline may operate and, as space that is not specifically dedicated to any single airline. In this model, all airline usable airport space is available to be used by any airline. The goal of the full “Common Use Model” is to minimize the amount of time any given airline resource is not in use, as well as maximize the full use of the airport. Airports benefit from increased utilization of existing resources. In a full common use airport, airlines are assigned with no preferences given to any individual airline, similar to the air traffic control process. To manage resources properly, computer software and systems are put in place to perform complex calculations, monitor usage, and provide status reporting. In the “Common Use Model”, the ground handling (GH) companies are the key point because of their central role; the GH companies have to provide service to different types of incoming flights, airlines, passenger, cargo services, etc., while working in a dynamic, and safety restricted shared environment. These companies need to reduce costs and offer accurate results, to the service contract with the airlines, to avoid been charged for delays. In most of cases, the GH companies are directly contracted by the airlines in a rigid contract to assist with supplies. The way to reduce cost is the optimization of the assistant resources to the incoming flights. The optimization is focused in the reduction of the number of resources and the time to assist a flight, what means an increase on productivity.
The analyzed technologies of communication between readers were WIFI, WIMAX, and Zigbee. The reasons for selecting Zigbee were its characteristics of acceptable distances, auto routing capability, low consumption and a sufficient bandwidth. Zigbee is a specification of a set of high-level protocols for wireless communication, for use with low-power digital radios, based on IEEE 802.15.4 WPAN (Wireless Personal Area Network). These applications require secure communication with low data transmission rates and maximization of battery life. Its main features are the low consumption, the mesh network topology and its easy integration (nodes can be manufactured with very little electronics). Zigbee network topologies allow three alternatives: star, tree and mesh. For the situation in hand, the mesh topology is particularly interesting, allowing to maintain the communications even in the event of failure in one or more of the Zigbee nodes as new routes can be recalculated by the coordinator node. In the tested Zigbee system, the communication module can cover around 400 meters with a rate of 250 Kb/s, performing a large number of identifications per second between nodes. In this option, the bottleneck is the read speed of the RFID reader because of the implemented anti-collision protocol. RFID is the technology of choice for product identification and there is a multi-level standard available: the EPC (Electronic Product Code). The proposed end nodes can cost around 12-15 $, with an acceptable reading rate of around 1012 meters. The objective of this distance is to locate the position within an area, for example, in a specific door or in a defined point. Looking at the enterprise point of view, the use of EPC permits a transparent way to share the information of this WSN between the agents involved, which is a key factor to be addressed in industrial environments. In the case of airports, information has to be shared in real time between ground handling companies, airlines, airport suppliers and air traffic control. This means that these company agents can access the WSN because they share a common defined data model.
2. THE UNION OF WSN & RFID Looking at the problems presented in the previous chapter; this paper proposes a mix of the technologies RFID and WSN to track resources in industrial environments, such as an airport, a factory or a logistic centre. The identified resources can be vehicles, products, pallets, machinery and employers. The idea is based on having a wireless communication network able to cover big distances on specific zones inside the industrial environment through RFID.
Fig. 1. Communication Architecture. By using WSN based on Zigbee with RFID reader nodes, the identification process can cover big distances and, at the same time, focus in localized points with a common data
model. The communication architecture is depicted in the Figure 1. The benefits of this mix of technology are: •
The information flows around the dynamic nodes to the gateway or to the coordinator of the network in a transparent way.
•
The RFID identifies the sensorized elements through an EPC, which is a standard for the identification and communication protocols.
•
The WSN allows covering big distances in a dynamic way and RFID focuses the identifications in specific zones.
•
The cost of the RFID end nodes (10-12 $) is low compared with other existing devices.
•
The battery consumption of RFID tag devices in the mobile nodes is really low compared with other electronic devices like mobile phones.
•
The possibility of auto routing of packages in case of node failure in the WSN is guaranteed by using Zigbee. 3. PROPOSED WSN IN THREE LEVELS: ZIGID
Using the proposed technologies, this chapter presents a WSN architecture divided in three levels of elements. The division is based in an adaptation of the adequate hardware devices to each application. The first level, or mobile-device level, is the end node level, the features of these end nodes are low consumptions, integrated sensors, short cover distances and identification process. The objective of the second level is the communication and routing of the information provided by the end nodes; this is the routing level. The third and last level is the coordination level, which has the function of coordination and provides an interface with the existing information systems (Fig. 2).
This level is the end node level; the hardware devices on this level are attached to the resources to be tracked, thus being mobile devices. The functionality of this level is to communicate the identification of the resource to the WSN through an active RFID communication. At this level, there are some limitations: the use of batteries, the computation capacity and the communication bandwidth. The proposed devices are active tags with sensors; which means that the tag needs to have batteries. The elements that compose the device being used in this example are, basically, a microcontroller and a transceiver. The microcontroller is a PIC16F819 made by Microchip. The main feature of the PIC16F819 is its low power consumption (7µA); as it integrates the use of nanoWatt Technology. The microcontroller is in charge of managing the operations of the end node: communications with routers and data acquisition from sensors. The transceiver is the nRF905 made by Nordic. It makes possible the communication between nodes and routers. NRF905 is a RF transmitter/receiver that operates in the UHF band. It is configurable to work at 433, 868 and 915 MHz. It also has different transmission power modes from -10 dBm to 10 dBm, allowing for a high read range with low consumption. It is commanded through a Synchronous Serial Port, SPI. Other elements of the node are the antenna, sensors and batteries. The antenna is a dipole printed on the board, of the type known as patch. The node can operate different sensors to measure temperature, acceleration (shocks), fuel and chemicals detectors among others. Two conventional AAA batteries provide the required power supply. Some tests are being carried out to calculate the autonomy of the end node. It is estimated to last an overall of 2844,95 hours, for a duty cycle where there is a RF communication every 2 seconds with a duration around 0.1 seconds in each transfer. This measurement has been realized at 10 dBm, which is the higher power the device can manage. It is necessary to optimise the communication protocol in order to increase the autonomy. 3.2 Router level The second level takes in charge of the routing of the information to the coordinator; this level has two communication protocols. The first one gets the information from the end nodes through RFID. The second one routes the information to the network coordinator using Zigbee.
Fig. 2. ZigID Architecture. 3.1 End node level
Zigbee allows auto routing of nodes and the information flows to the coordinator. Sometimes, the management staff needs to improve the visibility in real time; therefore, it is necessary to be able to add new devices in the WSN on a dynamic way. This requirement calls for a specialization in the software, which will be explained in the next chapter. Other functionality of this device is the filtering of to filter the unnecessary information, as redundant identifications or
non well-defined information, in order to avoid blocking the communication with a massive traffic. Since these networks are very critical in industrial environments, the design of the Zigbee network must always count on the possibility of different routing paths, especially since these devices are powered with batteries. This was one of the main concerns expressed by the end users. The functionality of this level is the communication with the end nodes by RFID and the routing of the packages in the WSN through Zigbee. The main elements constituting the router are the following: a JN5139 by Jennic and an nrF905. As in the previous case, the nRF905 is in charge of the RFID communication. The JN5139 is a module that enables the implementation of Zigbee networks. There are two types of these Jennic modules. For this development, the high power one has been used. It can transmit with a maximum power of 19 dBm, up to 4 km according to the IC data-sheet, although we have achieved read ranges of about 400m in a real environment. JN5139 also includes microcontroller features, allowing for the control of serial communications and input/output pins. This way, JN5139 manages the Zigbee network, the RFID communications and some buttons and leds. Routers are outdoor devices placed at fixed locations. The best option is to plug them in the power grid. In isolated places, the use of high-capacity rechargeable batteries with alternative power sources, like photocells, would cope with the autonomy problem. On this paper, the improvement of energy management of the batteries has been made in order to increase battery-life. 3.3 Coordinator Level The coordinator level has the objective of connecting the WSN with the information systems. The information captured by the WSN becomes adequately formatted in XML and/or is submitted in a query to a database. During the implementation of this prototype, the coordinator submitted all the information to a Mysql database, which was accessible to the rest of the systems of the airport. There is only one coordinator device in the WSN; which is called Pan Coordinator. It selects the frequency channel to be used by the network (usually the one with the least detected activity); then, it starts the network and, finally, it allows child nodes to connect to it (that is, to join the network). It can also provide message routing, security management and other services. The prototype has been developed to allow direct USB communication with a personal computer; future developments may also include Ethernet communication. 3. THE INFORMATION FLOW ON ZIGID Based on the WSN describe before, a data model needs to be defined in order to describe an information flow that satisfies the global requirements. The data model defines the object event as the main data shared object in the network, while
defining the management event as internal information to maintain the facilities on the network. The object event is serialized in the router node and it is communicated to the rest of the WSN. The steps of the data in the WSN are defined in the next figure as: 1.
The end node has the stored information in the local memory while its sensors are taking data from the environment.
2.
The reader router node receives the information through RFID.
3.
The router node prepares the object event.
4.
The sent object event is routed in the WSN.
5.
The Coordinator receives the object event and sends the data to the existing information systems.
Fig. 3. Information flow on ZigID. The object event is a software structure that represents the information definition step, which may includes far more information than a simple identification. When the reader router node establishes a communication with an end node this one sends the identification EPC (Electronic Product Code), the value of the variables about the sensors and a timestamp. After that, the reader router node, with the information from the end node, composes the object event adding also information about itself, like the location of the reader, the ID of the reader or the business process. Then, this object event is injected in the WSN. When the package arrives to the coordinator, it defines a XML file to transmit the data to the information system. During the demo, the XML generates a query that is submitted to a Mysql Data base. The next step will be a dynamic connection to Google Earth. The information about the nodes will be shown in the Google Earth tool as a layer of resources; the tool will present information about the position and the information represented in the objects events. The other type of object defined in the data model is a management package that contains status information about the end nodes and router nodes: the state of the batteries, problems with the antennas, routing problems, fallen devices or Zigbee exceptions. Many of these exceptions are caught by the functionality of the Zigbee stack. With a proper definition of this stack, the coordinator will tackle many problems in the network; generating alarms or specific recovery events.
4. APPLICATION IN GROUND HANDLING Different researchers and companies have developed real industrial applications of WSN, with satisfactory results (Sung, Lopez, & Kim, 2007) (Garcia, Garcia & Pastor, 2008). In the case at hand, the goal of this Project is to develop a WSN to improve the visibility in the terminal to help with resource management at the new Ciudad Real Central Airport (Figure 4). This airport entered in operation at the beginning of 2008 at Ciudad Real (Spain). The airport focuses in the low cost market; which means that it needs to reduce costs in ground handling operations, in airline management and other airport operations. The design of the airport is based in a hub with four aircraft parking positions facing the passenger gates. Close to the main parking there is a secondary bigger area for long time parking.
than 400 meters. Furthermore, there are defined mirror paths in all the routes.
Fig. 5. WSN Model for the airport. Besides the management information, the WSN can also offer information such as:
Fig. 4. Ciudad Real Central Airport. Trying to improve operations, the proposed solution tool allocates resources to incoming flights through the WSN. The data thus obtained is presented in a user interface that can run/show the state of the resources in different ways. This visualization tool helps airport management to identify possible conflict situations, thus improving the decision making process. As was said in the introduction, the formatted information coming from the WSN will be used in an artificial intelligent process. At a first approach the airport staff wants the Multiagent System technology to support decision making in real time. In order to improve short-term planning, the business expert system is permanently updated with the data from the WSN. Before the real implementation, a demo of the application, based on the installations of the airport, has been carried out. This first validation step includes end nodes in the majority of resources and selected positions for the router nodes. To begin with, the airport staff wanted to know if the resources were free or busy. The defined positions for busy resources were the four-parking-spaces hub in front of the gates plus the ”resources in transit”. The resources are considered free when they are in the hangars for ground handling resources. The idea of the association of resources with their involved process has also been analyzed. The system can generate alarms or re-calculate the planning when there are delays in the expected forecast plan. The first topology of WSN is shown in Figure 5 where the end nodes are being carried by the mobile ground-handling resources. The router nodes are situated in the principal parking positions and the hangars of resources; with this deployment, the distances between routers is always shorter
•
Administration information: the system can control working hours, productivity, statistics of performance, logs, identification of errors, etc.
•
Security information: the WSN can generate alarms when the tracked resources are situated in security zones or dangerous places such as the taxi routes for airplanes.
Fig. 6. Visual software tool. Figure 6 shows the interface of the application during the demo realized to the airport management staff. On the left, the interface shows the time information and the free resources that are in the hangar. The application has a player with a time bar; which can be taken forward or backward to get the historical representation of the information captured by the WSN. On the right, there are four pictures with the incoming aircrafts and all the resources that are attending the flights. The identified resources by the WSN are shown in this interface around the aircraft. The airport staff provided a very positive feedback in which they especially appreciated that, with a brief look at the interface; they could get a well-
formed picture of the state of the ground handling resources in the airport. 6. CONCLUSIONS The result of this work is a real implementation of a WSN based on RFID as identification technology and Zigbee as communication technology. The existing developed standards abstract the implementation of the low layers of the WSN, focusing the work in the software deployment, the data model and the requirements of the final user. The aim of the WSN is to cover big distances in the environment focusing the identification in defined zones with a flexible data model that allows adding new devices or new information on line. The proposed WSN is based on a three-level architecture made up by an end node level (RFID), a router level (RFID + Zigbee) and a coordinator level (Zigbee + Database). Due to the requirements of this outdoor mobile application an energy efficient management was necessary. This work proposes different software techniques based on the power adaptability, ultra low power consumption modes and efficiency routing; getting high performance consumption ratios. Complementing the WSN, a visual software tool has been develop to manage the ground handling resources that assist the incoming flight in the Ciudad Real Central Airport. The airport staff had a very positive reaction accepting the application willingly once they got familiar with the demo. This was mainly due to the ease of use, the simplicity and directness in which visual information was provided. REFERENCES Akkaya, K., & Younis, M. (2005). A survey on routing protocols for wireless sensor networks. Ad Hoc Networks, 3, 325-349. Akyildiz, I., Su, W., Sankarasubramaniam, Y., & Cayirci, E. (2002). Wireless sensor networks: a survey. Computer Networks, 38, 393-422. Thorne A., D. B., McFarlane D.. (2007). Impact of RFID on Aircraft Operations Processes. White Papers. Aero-ID. Cambridge. Bussman S., J. N. R. a. W. M. (2004). Multiagent System for Manufacturing Control.A Design Methodology. Springer Series on Agent Technology. Springer-Verlag: Berlin, Germany. Chen, J., Kher, S., & Somani, A. (2006). Distributed fault detection of wireless sensor networks. In (pp. 72): ACM. E.H., D. (1996). Planning in Distributed Artificial Intelligence. Foundations of Distributed Artificial Intelligence, pp. 187 – 210. John Wiley & Sons: New York, NY, USA. Garcia A., M. D., Thorne A., Fletcher M. (2003). The Impact of Auto-ID Technology in Material Handling Systems. 7th IFAC conference of Intelligent Manufacturing Systems (IMS’03), Budapest, Hungary. Garcia Ansola P. Garcia A., Pastor J.M. (2008). RFID in Physical Platforms of Agents: Application in Airport Management. Flins 2008. Madrid. Ghiasi, S., Srivastava, A., Yang, X., & Sarrafzadeh, M.
(2002). Optimal energy aware clustering in sensor networks. Sensors, 2, 258-269. Grossberg, S. (1988). Nonlinear neural networks: Principles, mechanisms, and architectures. Neural Networks, 1, 1761. Heinzelman, W., Kulik, J., & Balakrishnan, H. (1999). Adaptive protocols for information dissemination in wireless sensor networks. In (pp. 174-185): ACM New York, NY, USA. Jensen, F. (1996). An introduction to Bayesian Networks. UCL press London. Klir, G., & Yuan, B. (1995). Fuzzy sets and fuzzy logic: theory and applications: Prentice Hall Upper Saddle River, NJ. Russell, S., & Norvig, P. (1995). Artificial intelligence: a modern approach. New Jersey. Sohrabi, K., Gao, J., Ailawadhi, V., & Pottie, G. (2000). Protocols for self-organization of a wireless sensor network. IEEE personal communications, 7, 16-27. Ssu, K., Chou, C., Jiau, H., & Hu, W. (2006). Detection and diagnosis of data inconsistency failures in wireless sensor networks. Computer Networks, 50, 1247-1260. Sung, J., Lopez, T., & Kim, D. (2007). The EPC Sensor Network for RFID and WSN Integration Infrastructure. In (pp. 618-621): IEEE Computer Society. T.W., S. (1999). Distributed Rational Decision Making. Multi-Agent Systems, pp. 201 – 258, MIT Press: Cambridge, MA, USA.
Escuela Técnica Superior de Ingenieros Industriales, Universidad de Castilla-La Mancha, Ciudad Real, 13071, Spain. ([email protected]). 2 Universidad Rey Juan Carlos, Facultad de Ciencias Jurídicas y Sociales, Paseo Artilleros, Madrid, 28032, Spain. 3 Departamento de Electrónica, Universidad de Jaén, EPS, Campus las Lagunillas, Jaén, 23071, Spain.
Abstract: The new economic situation in all industries, and especially in the air transportation industry, claims for new business models supported by accurate management processes, which need constant feedback of the real status of the environment. A combined use of identification technologies and wireless sensor networks (WSN) can provide the required visibility to that information on status. This work proposes a WSN based on a Zigbee network of RFID readers with active tags as end nodes. The WSN can get the identification (ID) information stored in the end nodes, which are carried by the physical resources. Besides the ID, the tag can include other useful data coming from sensors such as acceleration, temperature, humidity, fuel level, etc. The final objective is to achieve an updated/real picture of the resources in the airport environment that will help in the decision-making. Keywords: Computer Communication Networks, RFID, Intelligent Manufacturing System.
1. INTRODUCTION New management systems are steadily increasing their complexity due to pressing requirements for productivity and flexibility. This complexity becomes especially poignant when dealing with typically big plants, airports or logistic centers; where the usual control is too rigid and hard to manage, maintain and update (Thorne A., 2007). For a modern system however, finding an adequate response to a broad range of possible new situations must be business-asusual; and this should apply even while the system is being modified/updated (Garcia A., 2003). Thus, distributed control becomes an standard practice as it allows a division of the problem in smaller and more affordable ones. In this case the system becomes distributed; where a negotiation process is relied upon, to take decisions that affect more than one of these parts (E.H., 1996). The need for an improvement in the connection between the real environment and the information systems is essential (T.W., 1999); ERPs (Enterprise Resource Planning) or any other information/management system has given place to new architectures, which need constant feed back of information of the real environment. This information is managed by new artificial intelligent techniques like the Bayesian networks (Jensen, 1996), Multiagent systems (Bussman S., 2004), Neural networks (Grossberg, 1988) or Fuzzy systems (Klir & Yuan, 1995); making expert system adapted to the environment (Russell & Norvig, 1995). To be able to get information from the environment, the systems need a transparent identification technology and quick communications to the information system. RFID technology is perfect to capture information from the physical world; offering rapid responses and continuous tracking. RFID
stands for radio frequency identification; a technology in which there are standards already in use defining communication protocols, hardware and information systems. An example of these standards is the one provided by EPC (Electronic Product Code). However, the reading range of RFID systems is relatively short for some applications, typically from two to three meters for passive tags in UHF and about 8-12 meters for active tags. As known, there are no standards available for a topology composed by several readers as the one typically used for WSN (Akyildiz, Su, Sankarasubramaniam, & Cayirci, 2002). The WSN abstracts the mobility, the topology, the scalability and the communication to the information system (Sohrabi, Gao, Ailawadhi, & Pottie, 2000); besides, the routing information can be added to the information package submitted (Akkaya & Younis, 2005). From this point of view, identification is another way of sensing. The ID is attached to the information of the sensor, getting a complete information package made up of the variables of sensors, the routing path and the unique ID of the sensorized element; making a complete data model in the WSN (Heinzelman, Kulik, & Balakrishnan, 1999). One of the main requirements in the new management is the needed necessity of a dynamic visibility of the environment. The staff does not know where they will need to improve the visibility. It means that the proposed WSN needs to be dynamic in the topology of the network and its information submitted. The second main problem is the needed of a lowlevel consumption in the WSN (Ghiasi, Srivastava, Yang, & Sarrafzadeh, 2002) and the detection (Chen, Kher, & Somani, 2006)& auto diagnostic system (Ssu, Chou, Jiau, & Hu, 2006). The paper will focus on these problems in Chapter 2.
The proposed WSN will be implemented in the Ciudad Real Central Airport in Spain. This is a small airport and its demand is based on low cost companies and cargo operations. The airport wants to improve the productivity of the resources of ground handling trying to reduce delays, overloads and overstaff. That will allow for a reduction of the costs, following the path set by the low cost companies. 1.1 Current situation in airport operations The world economic crisis has forced airlines to modify the way in which they are operating: improving productivity and reducing costs. Modern airports can be extremely complex organizations, which often involve governmental organizations, private companies, airlines, aircraft operations and airport operators. New ideas like common use are exploiting to reduce cost and improve the productivity. The concept of “Common Use Model” indicates that airport operators can gain centralized control over facilities and technologies; increase passenger-processing options; and acquire shared use efficiencies as they move from exclusive use toward common use. Airport common usable space is defined as the space in which any airline may operate and, as space that is not specifically dedicated to any single airline. In this model, all airline usable airport space is available to be used by any airline. The goal of the full “Common Use Model” is to minimize the amount of time any given airline resource is not in use, as well as maximize the full use of the airport. Airports benefit from increased utilization of existing resources. In a full common use airport, airlines are assigned with no preferences given to any individual airline, similar to the air traffic control process. To manage resources properly, computer software and systems are put in place to perform complex calculations, monitor usage, and provide status reporting. In the “Common Use Model”, the ground handling (GH) companies are the key point because of their central role; the GH companies have to provide service to different types of incoming flights, airlines, passenger, cargo services, etc., while working in a dynamic, and safety restricted shared environment. These companies need to reduce costs and offer accurate results, to the service contract with the airlines, to avoid been charged for delays. In most of cases, the GH companies are directly contracted by the airlines in a rigid contract to assist with supplies. The way to reduce cost is the optimization of the assistant resources to the incoming flights. The optimization is focused in the reduction of the number of resources and the time to assist a flight, what means an increase on productivity.
The analyzed technologies of communication between readers were WIFI, WIMAX, and Zigbee. The reasons for selecting Zigbee were its characteristics of acceptable distances, auto routing capability, low consumption and a sufficient bandwidth. Zigbee is a specification of a set of high-level protocols for wireless communication, for use with low-power digital radios, based on IEEE 802.15.4 WPAN (Wireless Personal Area Network). These applications require secure communication with low data transmission rates and maximization of battery life. Its main features are the low consumption, the mesh network topology and its easy integration (nodes can be manufactured with very little electronics). Zigbee network topologies allow three alternatives: star, tree and mesh. For the situation in hand, the mesh topology is particularly interesting, allowing to maintain the communications even in the event of failure in one or more of the Zigbee nodes as new routes can be recalculated by the coordinator node. In the tested Zigbee system, the communication module can cover around 400 meters with a rate of 250 Kb/s, performing a large number of identifications per second between nodes. In this option, the bottleneck is the read speed of the RFID reader because of the implemented anti-collision protocol. RFID is the technology of choice for product identification and there is a multi-level standard available: the EPC (Electronic Product Code). The proposed end nodes can cost around 12-15 $, with an acceptable reading rate of around 1012 meters. The objective of this distance is to locate the position within an area, for example, in a specific door or in a defined point. Looking at the enterprise point of view, the use of EPC permits a transparent way to share the information of this WSN between the agents involved, which is a key factor to be addressed in industrial environments. In the case of airports, information has to be shared in real time between ground handling companies, airlines, airport suppliers and air traffic control. This means that these company agents can access the WSN because they share a common defined data model.
2. THE UNION OF WSN & RFID Looking at the problems presented in the previous chapter; this paper proposes a mix of the technologies RFID and WSN to track resources in industrial environments, such as an airport, a factory or a logistic centre. The identified resources can be vehicles, products, pallets, machinery and employers. The idea is based on having a wireless communication network able to cover big distances on specific zones inside the industrial environment through RFID.
Fig. 1. Communication Architecture. By using WSN based on Zigbee with RFID reader nodes, the identification process can cover big distances and, at the same time, focus in localized points with a common data
model. The communication architecture is depicted in the Figure 1. The benefits of this mix of technology are: •
The information flows around the dynamic nodes to the gateway or to the coordinator of the network in a transparent way.
•
The RFID identifies the sensorized elements through an EPC, which is a standard for the identification and communication protocols.
•
The WSN allows covering big distances in a dynamic way and RFID focuses the identifications in specific zones.
•
The cost of the RFID end nodes (10-12 $) is low compared with other existing devices.
•
The battery consumption of RFID tag devices in the mobile nodes is really low compared with other electronic devices like mobile phones.
•
The possibility of auto routing of packages in case of node failure in the WSN is guaranteed by using Zigbee. 3. PROPOSED WSN IN THREE LEVELS: ZIGID
Using the proposed technologies, this chapter presents a WSN architecture divided in three levels of elements. The division is based in an adaptation of the adequate hardware devices to each application. The first level, or mobile-device level, is the end node level, the features of these end nodes are low consumptions, integrated sensors, short cover distances and identification process. The objective of the second level is the communication and routing of the information provided by the end nodes; this is the routing level. The third and last level is the coordination level, which has the function of coordination and provides an interface with the existing information systems (Fig. 2).
This level is the end node level; the hardware devices on this level are attached to the resources to be tracked, thus being mobile devices. The functionality of this level is to communicate the identification of the resource to the WSN through an active RFID communication. At this level, there are some limitations: the use of batteries, the computation capacity and the communication bandwidth. The proposed devices are active tags with sensors; which means that the tag needs to have batteries. The elements that compose the device being used in this example are, basically, a microcontroller and a transceiver. The microcontroller is a PIC16F819 made by Microchip. The main feature of the PIC16F819 is its low power consumption (7µA); as it integrates the use of nanoWatt Technology. The microcontroller is in charge of managing the operations of the end node: communications with routers and data acquisition from sensors. The transceiver is the nRF905 made by Nordic. It makes possible the communication between nodes and routers. NRF905 is a RF transmitter/receiver that operates in the UHF band. It is configurable to work at 433, 868 and 915 MHz. It also has different transmission power modes from -10 dBm to 10 dBm, allowing for a high read range with low consumption. It is commanded through a Synchronous Serial Port, SPI. Other elements of the node are the antenna, sensors and batteries. The antenna is a dipole printed on the board, of the type known as patch. The node can operate different sensors to measure temperature, acceleration (shocks), fuel and chemicals detectors among others. Two conventional AAA batteries provide the required power supply. Some tests are being carried out to calculate the autonomy of the end node. It is estimated to last an overall of 2844,95 hours, for a duty cycle where there is a RF communication every 2 seconds with a duration around 0.1 seconds in each transfer. This measurement has been realized at 10 dBm, which is the higher power the device can manage. It is necessary to optimise the communication protocol in order to increase the autonomy. 3.2 Router level The second level takes in charge of the routing of the information to the coordinator; this level has two communication protocols. The first one gets the information from the end nodes through RFID. The second one routes the information to the network coordinator using Zigbee.
Fig. 2. ZigID Architecture. 3.1 End node level
Zigbee allows auto routing of nodes and the information flows to the coordinator. Sometimes, the management staff needs to improve the visibility in real time; therefore, it is necessary to be able to add new devices in the WSN on a dynamic way. This requirement calls for a specialization in the software, which will be explained in the next chapter. Other functionality of this device is the filtering of to filter the unnecessary information, as redundant identifications or
non well-defined information, in order to avoid blocking the communication with a massive traffic. Since these networks are very critical in industrial environments, the design of the Zigbee network must always count on the possibility of different routing paths, especially since these devices are powered with batteries. This was one of the main concerns expressed by the end users. The functionality of this level is the communication with the end nodes by RFID and the routing of the packages in the WSN through Zigbee. The main elements constituting the router are the following: a JN5139 by Jennic and an nrF905. As in the previous case, the nRF905 is in charge of the RFID communication. The JN5139 is a module that enables the implementation of Zigbee networks. There are two types of these Jennic modules. For this development, the high power one has been used. It can transmit with a maximum power of 19 dBm, up to 4 km according to the IC data-sheet, although we have achieved read ranges of about 400m in a real environment. JN5139 also includes microcontroller features, allowing for the control of serial communications and input/output pins. This way, JN5139 manages the Zigbee network, the RFID communications and some buttons and leds. Routers are outdoor devices placed at fixed locations. The best option is to plug them in the power grid. In isolated places, the use of high-capacity rechargeable batteries with alternative power sources, like photocells, would cope with the autonomy problem. On this paper, the improvement of energy management of the batteries has been made in order to increase battery-life. 3.3 Coordinator Level The coordinator level has the objective of connecting the WSN with the information systems. The information captured by the WSN becomes adequately formatted in XML and/or is submitted in a query to a database. During the implementation of this prototype, the coordinator submitted all the information to a Mysql database, which was accessible to the rest of the systems of the airport. There is only one coordinator device in the WSN; which is called Pan Coordinator. It selects the frequency channel to be used by the network (usually the one with the least detected activity); then, it starts the network and, finally, it allows child nodes to connect to it (that is, to join the network). It can also provide message routing, security management and other services. The prototype has been developed to allow direct USB communication with a personal computer; future developments may also include Ethernet communication. 3. THE INFORMATION FLOW ON ZIGID Based on the WSN describe before, a data model needs to be defined in order to describe an information flow that satisfies the global requirements. The data model defines the object event as the main data shared object in the network, while
defining the management event as internal information to maintain the facilities on the network. The object event is serialized in the router node and it is communicated to the rest of the WSN. The steps of the data in the WSN are defined in the next figure as: 1.
The end node has the stored information in the local memory while its sensors are taking data from the environment.
2.
The reader router node receives the information through RFID.
3.
The router node prepares the object event.
4.
The sent object event is routed in the WSN.
5.
The Coordinator receives the object event and sends the data to the existing information systems.
Fig. 3. Information flow on ZigID. The object event is a software structure that represents the information definition step, which may includes far more information than a simple identification. When the reader router node establishes a communication with an end node this one sends the identification EPC (Electronic Product Code), the value of the variables about the sensors and a timestamp. After that, the reader router node, with the information from the end node, composes the object event adding also information about itself, like the location of the reader, the ID of the reader or the business process. Then, this object event is injected in the WSN. When the package arrives to the coordinator, it defines a XML file to transmit the data to the information system. During the demo, the XML generates a query that is submitted to a Mysql Data base. The next step will be a dynamic connection to Google Earth. The information about the nodes will be shown in the Google Earth tool as a layer of resources; the tool will present information about the position and the information represented in the objects events. The other type of object defined in the data model is a management package that contains status information about the end nodes and router nodes: the state of the batteries, problems with the antennas, routing problems, fallen devices or Zigbee exceptions. Many of these exceptions are caught by the functionality of the Zigbee stack. With a proper definition of this stack, the coordinator will tackle many problems in the network; generating alarms or specific recovery events.
4. APPLICATION IN GROUND HANDLING Different researchers and companies have developed real industrial applications of WSN, with satisfactory results (Sung, Lopez, & Kim, 2007) (Garcia, Garcia & Pastor, 2008). In the case at hand, the goal of this Project is to develop a WSN to improve the visibility in the terminal to help with resource management at the new Ciudad Real Central Airport (Figure 4). This airport entered in operation at the beginning of 2008 at Ciudad Real (Spain). The airport focuses in the low cost market; which means that it needs to reduce costs in ground handling operations, in airline management and other airport operations. The design of the airport is based in a hub with four aircraft parking positions facing the passenger gates. Close to the main parking there is a secondary bigger area for long time parking.
than 400 meters. Furthermore, there are defined mirror paths in all the routes.
Fig. 5. WSN Model for the airport. Besides the management information, the WSN can also offer information such as:
Fig. 4. Ciudad Real Central Airport. Trying to improve operations, the proposed solution tool allocates resources to incoming flights through the WSN. The data thus obtained is presented in a user interface that can run/show the state of the resources in different ways. This visualization tool helps airport management to identify possible conflict situations, thus improving the decision making process. As was said in the introduction, the formatted information coming from the WSN will be used in an artificial intelligent process. At a first approach the airport staff wants the Multiagent System technology to support decision making in real time. In order to improve short-term planning, the business expert system is permanently updated with the data from the WSN. Before the real implementation, a demo of the application, based on the installations of the airport, has been carried out. This first validation step includes end nodes in the majority of resources and selected positions for the router nodes. To begin with, the airport staff wanted to know if the resources were free or busy. The defined positions for busy resources were the four-parking-spaces hub in front of the gates plus the ”resources in transit”. The resources are considered free when they are in the hangars for ground handling resources. The idea of the association of resources with their involved process has also been analyzed. The system can generate alarms or re-calculate the planning when there are delays in the expected forecast plan. The first topology of WSN is shown in Figure 5 where the end nodes are being carried by the mobile ground-handling resources. The router nodes are situated in the principal parking positions and the hangars of resources; with this deployment, the distances between routers is always shorter
•
Administration information: the system can control working hours, productivity, statistics of performance, logs, identification of errors, etc.
•
Security information: the WSN can generate alarms when the tracked resources are situated in security zones or dangerous places such as the taxi routes for airplanes.
Fig. 6. Visual software tool. Figure 6 shows the interface of the application during the demo realized to the airport management staff. On the left, the interface shows the time information and the free resources that are in the hangar. The application has a player with a time bar; which can be taken forward or backward to get the historical representation of the information captured by the WSN. On the right, there are four pictures with the incoming aircrafts and all the resources that are attending the flights. The identified resources by the WSN are shown in this interface around the aircraft. The airport staff provided a very positive feedback in which they especially appreciated that, with a brief look at the interface; they could get a well-
formed picture of the state of the ground handling resources in the airport. 6. CONCLUSIONS The result of this work is a real implementation of a WSN based on RFID as identification technology and Zigbee as communication technology. The existing developed standards abstract the implementation of the low layers of the WSN, focusing the work in the software deployment, the data model and the requirements of the final user. The aim of the WSN is to cover big distances in the environment focusing the identification in defined zones with a flexible data model that allows adding new devices or new information on line. The proposed WSN is based on a three-level architecture made up by an end node level (RFID), a router level (RFID + Zigbee) and a coordinator level (Zigbee + Database). Due to the requirements of this outdoor mobile application an energy efficient management was necessary. This work proposes different software techniques based on the power adaptability, ultra low power consumption modes and efficiency routing; getting high performance consumption ratios. Complementing the WSN, a visual software tool has been develop to manage the ground handling resources that assist the incoming flight in the Ciudad Real Central Airport. The airport staff had a very positive reaction accepting the application willingly once they got familiar with the demo. This was mainly due to the ease of use, the simplicity and directness in which visual information was provided. REFERENCES Akkaya, K., & Younis, M. (2005). A survey on routing protocols for wireless sensor networks. Ad Hoc Networks, 3, 325-349. Akyildiz, I., Su, W., Sankarasubramaniam, Y., & Cayirci, E. (2002). Wireless sensor networks: a survey. Computer Networks, 38, 393-422. Thorne A., D. B., McFarlane D.. (2007). Impact of RFID on Aircraft Operations Processes. White Papers. Aero-ID. Cambridge. Bussman S., J. N. R. a. W. M. (2004). Multiagent System for Manufacturing Control.A Design Methodology. Springer Series on Agent Technology. Springer-Verlag: Berlin, Germany. Chen, J., Kher, S., & Somani, A. (2006). Distributed fault detection of wireless sensor networks. In (pp. 72): ACM. E.H., D. (1996). Planning in Distributed Artificial Intelligence. Foundations of Distributed Artificial Intelligence, pp. 187 – 210. John Wiley & Sons: New York, NY, USA. Garcia A., M. D., Thorne A., Fletcher M. (2003). The Impact of Auto-ID Technology in Material Handling Systems. 7th IFAC conference of Intelligent Manufacturing Systems (IMS’03), Budapest, Hungary. Garcia Ansola P. Garcia A., Pastor J.M. (2008). RFID in Physical Platforms of Agents: Application in Airport Management. Flins 2008. Madrid. Ghiasi, S., Srivastava, A., Yang, X., & Sarrafzadeh, M.
(2002). Optimal energy aware clustering in sensor networks. Sensors, 2, 258-269. Grossberg, S. (1988). Nonlinear neural networks: Principles, mechanisms, and architectures. Neural Networks, 1, 1761. Heinzelman, W., Kulik, J., & Balakrishnan, H. (1999). Adaptive protocols for information dissemination in wireless sensor networks. In (pp. 174-185): ACM New York, NY, USA. Jensen, F. (1996). An introduction to Bayesian Networks. UCL press London. Klir, G., & Yuan, B. (1995). Fuzzy sets and fuzzy logic: theory and applications: Prentice Hall Upper Saddle River, NJ. Russell, S., & Norvig, P. (1995). Artificial intelligence: a modern approach. New Jersey. Sohrabi, K., Gao, J., Ailawadhi, V., & Pottie, G. (2000). Protocols for self-organization of a wireless sensor network. IEEE personal communications, 7, 16-27. Ssu, K., Chou, C., Jiau, H., & Hu, W. (2006). Detection and diagnosis of data inconsistency failures in wireless sensor networks. Computer Networks, 50, 1247-1260. Sung, J., Lopez, T., & Kim, D. (2007). The EPC Sensor Network for RFID and WSN Integration Infrastructure. In (pp. 618-621): IEEE Computer Society. T.W., S. (1999). Distributed Rational Decision Making. Multi-Agent Systems, pp. 201 – 258, MIT Press: Cambridge, MA, USA.