Wireless sensor network

A sensor network (of English. Wireless sensor network) is a computer network of sensor nodes, tiny ( " mote " ) to relatively large ( " shoebox " ) communicating via radio computers in an infrastructure -based ( base stations) or in a self either work together organizing ad -hoc network, to query their environment by means of sensors and transmit the information. The targeted size of future sensor nodes made ​​known the idea under the heading Smart Dust (English " smart dust ").

Sensor networks have been designed as a military early warning system for monitoring pipelines and borders. Modern research, however, sees it as a replacement for costly sensor arrays in vehicle, merchandise manager in warehouses and supervisor of natural areas for pollutants, forest fires and animal migrations; the possible applications are as varied as the available sensors ( cf. sensors by measurement ).

Sensor networks are always in the stage of development, practical applications, there is for experimental and demonstration purposes. Widespread sensor networks, there are for professional applications. The most famous sensor network is the weather stations from different vendors, but the crosslinking is performed by conventional telecommunications networks. Comparable networks of actors are not known, because the energies required for actuators and protection against malfunctions provide considerably higher demands on the network and the nodes.

The smallest existing sensor nodes has a diameter of one millimeter (as of 2007), which revealed hitherto largest sensor network with sensor nodes about 1000 a great 1300 to 300 meters free land area (as at Dec 2004 ).

• 3.1 Tasks of the network protocols
• 3.2 Special Sensor Network Protocols 3.2.1 Media Access Protocols
• 3.2.2 routing protocols
• 3.2.3 Protocol Stacks

• 5.1 Synchronization by calculating the round trip time
• 5.2 Reference broadcast synchronization
• 5.3 Timing - Sync Protocol for Sensor Networks

• 6.1 Tiny Aggregation ( TAG)
• 6.2 Empirical mutual coding

History

As a precursor to the modern sensor network research, the Sound Surveillance System ( SOSUS ) can be considered an installed during the Cold War by the United States network of underwater buoys, the submarines tracking by acoustic sensors. Although SOSUS is not a computer network, however, brought the idea of a comprehensive sensor assembly out.

Research on sensor networks started around 1980 with the projects Distributed Sensor Networks (DSN ) and Sensor Information Technology ( Sensit ) of the Military Agency Defense Advanced Research Projects Agency ( DARPA ) of the United States. DARPA is working with military and university research institutions to develop new military and economically significant technologies. Your results are not subject to secrecy, what was true of the sensor network research generally.

In the 1990s, the sensor network research experienced a boom, encouraged by the ever smaller and more powerful expectant computer hardware. Today, sensor networks from research institutes around the world are examined. Results will be presented in 2003 at the " ACM Conference on Embedded Networked Sensor Systems " ( SenSys ).

Hardware

A sensor node is in the core as an ordinary computer, a processor and a data storage (typically flash memory). To one or more sensors and a module are to radio communications. All parts are powered by a battery with energy. In newer models, all components are housed on a single computer chip, which greatly reduces the size compared to composite components.

Some designs see network node before without sensors that are only for the communication and management. Do all nodes of a network have the same sensory equipment, so it is called a homogeneous sensor network, otherwise of a heterogeneous. Heterogeneous sensor networks are particularly useful if the sensors have different usage patterns, so for example in measuring frequency, measuring time and accumulating data volume are very different.

As a promising means of communication radio technology is currently traded it but other communication media such as light or sound were being considered. It is assumed that the communication module such as a wireless device is ready to send two states and listening know between which can be switched with little delay.

Sensor nodes obtained after the application of any new energy reserves; the reserves are depleted, the lifetime of the node is also exhausted. The battery of a sensor node must therefore be as efficient as possible, while all other parts must have the minimum possible power consumption. To reduce the power consumption further, each sensor node may be placed in a standby state, in which all parts are switched off until the internal processor clock. When disconnecting and turning you pronounce " go to sleep " and of "wake up", thus producing genuine " daily routines " come about. Abstaining from batteries through use of renewable energy such as photovoltaics would be desirable is the present state of the art but not feasible.

Future Music is also the desire for ecological sustainability: In the wild, expelled sensor nodes should rot without residue after the exhaustion of its energy reserves without polluting the environment, and animals should suffer through accidental ingestion of a sensor node is no damage.

The cost of the hardware should be so low that sensor networks on a large scale - the DARPA speaks of hundreds of thousands of sensor nodes - also financially viable. Sensor node with a unit price of € 2 would be justifiable in vehicle, large-scale applications for landscape monitoring with several thousand nodes are only viable on a price in the lower € - cent range; Today's sensor nodes reach the lower price range rarely and usually cost more than 100 € a piece.

Existing systems

In the past, some of the sensor node have been developed which are used to test specifically designed software and communication schemes. The nodes vary greatly in size, features and price, as they follow different targets: While some developers are trying to make their sensor nodes as small and cheap, while others fold versatility and ease of use for scientific testing applications. Known systems are:

• BTnode. Sensor node platform BTnodes that is custom enhanced sensors. The current model has BTnode rev3 the dimensions of 58.15 x 32.5 mm.
• EyesIFX
• FireFly FireFly by
• IDwaRF, iDwaRF -328 and iDwaRF box. Radio modules on Atmel AVR based for easy setup of wireless multipoint point-to- point (N: 1) networks in the 2.4GHz ISM band.
• Imote, Mica and Telos. Sensor node platforms from Crossbow Technology, which are custom enhanced sensors or come with standard equipment. The current models have the dimensions 36 x 48 x 9 mm ( Imote2 ), 58 x 32 x 7 mm ( Mica2 ) and 65 x 31 x 6 mm ( Telos B).
• INGA ( Inexpensive Node for General Applications). Open hardware sensor nodes developed by the Institute of Operating Systems and Computer Networks at the Technical University Braunschweig.
• Inode (intelligent Network Operating Device) sensor node platform of the Forschungszentrum Jülich in Flex PCB design. 20 x 20 x 5 mm (folded)
• ISense is a modular sensor network platform coalesenses. In addition to a basic module with processor and radio interface, there are different sensor modules (acceleration sensor and passive infrared, temperature and brightness, magnetic sensor ), power modules, and gateway module.
• Particles. Sensor nodes of the TecO the University of Karlsruhe with temperature, light and accelerometer. The current model has ³ is a size of less than 10 mm.
• Preon32, new sensor nodes to virtual machine
• Rene
• ScatterWeb
• S-net. Extremely energy -efficient wireless sensor networks at the Fraunhofer Institute for Integrated Circuits.
• SNoW5. Expandable sensor nodes at the University of Würzburg.
• Sun SPOT. Sensor node platform of Project Sun SPOT.
• TinyNode 584 Expandable sensor node with temperature sensor Shock Fish SA. The flat node has the dimensions 30 x 40 mm.
• Tmote Sky. Sensor nodes with temperature, light and humidity sensor of Moteiv. The flat node has the dimensions 32 x 80 mm.
• Waspmote. modular sensor nodes with the possibility to affix various sensors as an extension.
• WeC
• Wisenet
• Z1. Sensor nodes with temperature and acceleration sensor of Zolertia.

Communication

Sensor networks are ad hoc networks, ie networks without fixed infrastructure between devices. Ad -hoc networks are meshed networks in which network nodes are connected to one or more neighbors. This results in a multi-hop communications, be passed on the message from node to node until they reach their destination.

Such networks are characterized by an unpredictable dynamic behavior, because unlike permanently installed computer networks, the network topology is uncertain: the number and locations of the network nodes and " line quality " can not be foreseen during the operation nodes can be added or fail without warning.

Tasks of network protocols

The communication between the sensor nodes with each other is a key area of current research. The goal is to find network protocols that transmit data as efficiently as possible and at the same time the energy reserves of the sensor nodes save by allowing long bedtimes and energy-intensive components such as the radio unit address as seldom as possible.

A complete network protocol defines the behavior of the nodes in four matters:

• The initialization is the phase in which the sensor nodes find each other after deployment and build by locating their neighbors, the network topology. A clean design of the network topology is crucial for later success in routing.
• Underground flow is defined as the change between waking and sleeping times of the nodes. Because while sleeping times save energy, but can become unreachable nodes, it is important to find a reasonable middle ground here.
• The communication scheme determines runs as a single data exchange between two sensor nodes. It must be ensured that the data to be transmitted quickly and accurately, and not interfere with each other node.
• Finally, the routing specifies how messages are routed through the sensor network. Not always is here the shortest way is the best, because this could lead to a one-sided network load and thus premature failure of critical connection node. Previous research has walked these challenges usually separated and left it to the operator of the sensor network, to issue from the items a suitable method.

Sensor networks are particularly vulnerable to the classic problems of communication in computer networks, on the one hand, because a large number of terminals sharing a common communication medium to another, because sensor nodes are more affected by waste of resources affected as devices with a power outlet or rechargeable energy storage. For sensor networks are therefore suitable only protocols that can effectively avoid these problems.

Special sensor network protocols

Back in the early investigations of the military became clear that conventional network protocols are not suitable for sensor networks. Even today's standards for wireless networks such as IEEE 802.11 or Carrier Sense Multiple Access shall be wasteful with the energy reserves of the terminals to or can be as Bluetooth is not transmitted to networks with many participants. In addition, to stand sensor networks in one important respect from other mobile ad hoc networks: While compete usually in a network many different applications or users to the shared resources, there is a sensor network, only a single, network-wide application that in a sense competes with itself. Since the overall objective of the application takes precedence over the equality of individual nodes, the concept of fairness in computer networks must be designed new here.

The research protocol developed and thus tested network protocols that are tailored specifically to the needs of sensor networks. She goes in different directions, without so far a uniform standard would have emerged. Some researchers argue that the application areas of sensor networks are so different that it will never give the protocol for sensor networks, but always a choice of protocols that are different degrees for different purposes. The most important sensor network protocols are presented.

Media access control protocols

A large group of sensor network protocols is devoted to the role of the Media Access Control (MAC, engl. " Media access control") sharing of the communication medium (air). A primary role is played by the reduction of energy consumption. This is in contrast to traditional wireless networks (WLAN, GSM), where it comes as completely utilize the available bandwidth of the medium and to distribute these at the same time fair.

The radio module is often the component of the sensor nodes, which consumes the most energy. The energy consumption is similarly high for the different actuator types Bert the radio module (wait for news, receive, send ). The radio module is to save energy, therefore, mostly off ( engl. Duty Cycling). The MAC protocol must therefore decide not only when data is sent, but also when the radio module or should be turned off. Two methods come of it in use: Random access with carrier testing and time division multiplexing.

The Random access with carrier testing different variants of the so-called Low -Power Listening (LPL ) are used. The idea of LPL is that the radio is set regularly for a short time to check whether the medium is busy. This is not the case, the radio module is again issued to save energy. If the medium is busy the radio remains enabled to send and receive messages. For the transmitter, this approach involves knowing the difficulty when must be sent to ensure that the receiver also listens. The simple approach is to transmit a preamble that is longer than that of the receiver Aufwachinterval (Berkeley Media Access Control (B- MAC) ). Alternatively, a long stream of repetitive packets can be sent (X- MAC, SpeckMAC ). Can save energy ( and bandwidth) of the transmitter to the receiver Aufwachzeitplan learn ( WiseMAC ). As an alternative to LPL can also opposing the low-power probing (LPP ) are used. Here, a short carrier ( Beacon ) is regularly sent indicating that the node for a short time ready to receive a message is (RI - MAC).

At the time division multiplexing ( Time Division Multiple Access ( TDMA) ) a schedule is created when which nodes transmit and receive. This allows for a low-energy data exchange. However, caused the creation and maintenance of the schedule and the required synchronization overhead. Protocols in this class are Sensor Media Access Control (S- MAC), timeout Media Access Control (T- MAC), Dozer, SCP - MAC, LMAC, DMAC, TRAMA. When Dozer and DMAC note applies that MAC and routing are combined in a log.

Hybrid protocols like Crankshaft, Zebra Media Access Control (Z- MAC) or MAC SRTST try the advantages of random access with carrier testing and TDMA to combine.

Routing protocols

Routing protocols focus primarily the routing, so the question will be piloted as news quickly as possible and with as little effort on their destination. Network protocols that ignore the question of the routing, usually go from standard methods that are based on routing tables (see Routing). In fact, many routing protocols can be transmitted without or with only slight adjustments to sensor networks.

Of particular importance for sensor networks geographic routing processes are. In many application scenarios, the user is interested specifically for measurement data of a specific geographical area or point. Firstly, there are requests by the kind of " Deliver me all the data of the area xyz" with statements such as " On the node that is closest to the position xy " addressed to other nodes are. The network protocol must remove the user in the task of making the affected nodes detect and forward messages to them.

The process of Geo -Cast searches through a sensor network, all nodes in a selected geographical area out. By fitting and cut geometric shapes on a map the affected sensor nodes can be quickly identified. At the same time the user gets an easy-to -use graphical user interface.

A central role in routing plays the sensor network protocol Greedy Perimeter Stateless Routing in Wireless Networks ( GPSR ), the messages do not forward it to name but on geographical coordinates. These changes repeatedly between a greedy strategy, are passed in the data packets on a straight path toward the target, and a perimeter mode in which the data packet encircles the target. The perimeter mode to ensure that packets do not get stuck in dead ends in unfavorable network topologies. Geographic Hash Tables extend GPSR to the ability to distribute information to multiple neighboring nodes, thereby ensuring data security in case of failure of some nodes.

The routing protocol for sensor networks, which is being developed by the IETF RPL is.

Protocol Stacks

At present there are several competing protocol stacks of different consortia and organizations. Depending on the stack some or all layers of the OSI model are covered:

• ZigBee ( ZigBee Alliance)

• IPSO ( IPSO Alliance)

• Constrained Application Protocol (mainly IETF)

• IEEE 1451 ( IEEE)

• WirelessHART ( HART Communication Foundation )

• 6LoWPAN (IETF)

• NanoIP (IEEE, IETF)

Localization and tracking

Certain application scenarios and communication protocols require that a sensor node can determine its own location ( localization ) or the places of origin of measured signals ( localization). Since both issues are covered in other areas such as navigation and Astrophysics for centuries, today there are a variety of methods for a variety of initial conditions. However, it is to consider what processes can be performed with the limited technology of sensor networks and how the resulting work appropriately among the nodes of the network. So the widespread satellite location determination via Global Positioning System (GPS) is not suitable for sensor nodes, for example, as the necessary technical components have to be big, heavy and expensive.

Is there at least two sensor nodes in a sensor network that know their own position in the absolute geographical coordinates, and it is possible to gauge the distance between two sensor nodes to each other, it can determine its position in the geographic coordinate system, in general, each node of the network. The idea is that initially each sensor node establishes a personal coordinate system, with himself at the origin and two adjacent nodes as a guide for the x- and y- coordinate axes. Through methods such as triangulation, each node assigns all were neighbors in his personal system. Thereafter, the individual systems are combined by rotation and shift to a full coordinate system. If the network is arranged sparsely occupied or unfavorable, so the localization is inaccurate.

Can any two nodes determine their absolute position or can not be estimated distances, the localization is incomplete or inaccurate. For example, knows absolutely no sensor node 's absolute coordinates, so the sensor network can indeed mapped geographically correct under certain circumstances, but are not provided in the larger context of world coordinates. Missing the other hand, the possibility of distance measurement, so drop off triangulation and similar methods and the positions can be guessed only as cut faces several transmission radii. If both initial conditions are not met, the network topology can be only abstractly as a graph or an equivalent representation shown ( eg as adjacency matrix or incidence matrix ).

A comparatively recent approach is the determination of position by fingerprints (English fingerprints). Here, a node is created by listening to the radio channel a "fingerprint" called profile of the background noise. The background noise is influenced by the environment, such as near electric lines or walls on which radio waves are reflected, and differs from place to place. By comparison with a fingerprint database, a sensor node to estimate its own position. In this approach, prior knowledge about the application area or an additional auxiliary system are necessary.

In order to locate the origin of a measured signal clearly, the signal of at least three sensor node must have been received. From the different signal propagation times to the sensor node, the signal origin can be determined exactly via Hyperbelortung. If the signal is received by only two sensor nodes or less, no clear detection is possible and the origin can be limited only by cutting the transmission radii or the transmission radius alone.

Synchronization

Measurement data often depend on absolute times. In addition, some communication protocols, such as SMACS require an accurate synchronization of the sensor nodes.

Like other computer networks, one also has to contend with sensor networks with the typical inaccuracies in the synchronization. The factors which affect the synchronization, it is the transmission time, so the time required for the transmitter to be ready to transmit, the access time, that is the time required for the transmitter to store the data on the communication medium, the propagation speed, ie how fast the message from the sender to the recipient, and the time of receipt, so long as the receiver needs to be tapped a message from the media and make available the information for that application. Since the different communication methods among other factors mainly affect the access time, it is useful to make the decision depends on a synchronization method also from the communication process.

Synchronization by calculating the round trip time

One can determine the difference between two clocks in a computer network by subtracting the times of two computers from each other and it then again the round trip time is subtracted, which is caused by the messages exchanged between two computers in order to exchange information on their times inform. In practice, it is to the average of the time periods that were required for transmission of the request and its response. In this method, all the above factors except the propagation velocity affect. At a high variance of propagation velocity therefore this method is preferable.

Reference broadcast synchronization

In the Reference Broadcast Synchronization ( RBS) is sent from a central location from a synchronization signal to all nodes. A node that sends a message after this synchronization message, informed with this message at the same time the receiver about when he has received the synchronization. Using this information, the recipient can then decide if its clock is wrong, and they synchronize with the clock of the transmitter. This method is particularly well suited if in communicating the sending time and access times may vary, since the time at which is based synchronization, is sent only once for all recipients and different access broadcasting time so do not affect the synchronization.

Timing - Sync Protocol for Sensor Networks

Timing Sync Protocol for Sensor Networks ( TPSN ) describes the process of how synchronization is used in a sensor network. When synchronization method can be applied here, the synchronization by calculating the round trip time.

Initialization

Since there can still be collisions in spite of the random waiting time, there is also the possibility that node Request a level with level_request. These then are geared to the respective lowest level that they receive.

Synchronization

Aggregation

Some application scenarios require that data collected from the whole sensor network and sent finally to a single recipient ( " Central Valley "). Naive approaches to carrying out such an aggregation can lead to the formation of large amounts of data communication bottlenecks that hinder the performance of the system unnecessarily. Data aggregation and data compression help to avoid such bottlenecks. For example, if the maximum temperature measured in a sensor network are determined, it would be naive approach to transmit all measured temperatures at the Central Valley, which then picks out the maximum, while an advanced approach compares the data already in the transmission and ultimately only one single temperature value transmitted to the central Valley.

Tiny Aggregation ( TAG)

Tiny Aggregation ( TAG) treated the sensor network as a database are retrieved from the database using a query language data. Data requests are propagated from the Central Valley in a simple, SQL -like format in the sensor network. Nodes evaluate the request on their own data and received from neighbors and sort data so gradually redundant and unnecessary data in advance of.

Empirical mutual coding

The empirical mutual coding share information only if they do not meet the normal value. The principle is already being implemented in naive approaches, such as when a fire suppression sensor node forwards only those temperature measurements that exceed a preset value. The empirical mutual coding deepened this idea due to the observation that the measurements spatially close -spaced sensor nodes are always similar. A sensor node is its measurement relative to that of its neighbors, and only then, if the measured value differs greatly from that of the other. The strength of the approach is that the correlations of the measured values ​​are automatically determined.

Criticism of sensor networks

Michael Crichton recorded in 2002 in his novel prey a bleak future vision in which he combined the smart dust idea with collective intelligence and nanotechnology, and slew his fictional characters of swarms malicious microparticles.

Far more realistic criticism of sensor networks and the smart dust - thoughts to express privacy advocates. You see in sensor networks further monitoring method that can be abused for the monitoring of citizens and analysis of consumers without their knowledge or consent.