Localization in mobile communication is a broad topic that has received ... each with transceivers for wireless, short-range communication, capable of de-.
Evaluating Parameters for Localization in Wireless Sensor Networks: A survey Freddy L´ opez Villafuerte, Jochen Schiller, Ernesto Tapia Rodr´ıguez, Marte Ram´ırez Orteg´on, Erik Valdemar Cuevas∗ Freie Universit¨ at Berlin, Institut f¨ ur Informatik Takustr. 9, 14195 Berlin, Germany {lopez, schiller, tapia, ramirez}@inf.fu-berlin.de ∗ Universidad de Guadalajara, Depto. Electr´ onica y Computaci´ on Av. Revoluci´ on 1500, 44430 Guadalajara, Jalisco, M´exico {erik.cuevas}@cucei.udg.mx
Abstract. Localization in mobile communication is a broad topic that has received considerable attention from the research community during the past few decades. The approaches suggested to estimate location are implemented with different concepts, functionalities, scopes and technologies. The major contribution of this survey is to analyze and identify relevant parameters to compare different approaches and to point out the best method for low power and restricted memory wireless sensor nodes. Key words: Location techniques, Parameters for Localization, Signal technologies
1
Introduction
A Wireless Sensor Network (WSN) is formed by hundreds of small, low-cost nodes which have limitations in memory, energy, and processing capacity [1]. In this type of networks, one of the main problems is to locate each node. The vision of many researchers is to create smart environments, controlled through planned or ad-hoc deployment of a potentially large set of sensor nodes, each with transceivers for wireless, short-range communication, capable of detecting environment conditions such as temperature, movement, light, acoustic events or the presence of certain objects. WSN will enable fine-grained observation and control of the physical world. The futuristic scenario in sensor networks appears in large numbers of unattended autonomous nodes which operate in a dynamic environment. This kind of sensor will be able to organize itself. It will be aware of its physical position. The sensor nodes will carry out dynamic tasks in a distributed form, very frequently confronting change in the topology network and failures in the network nodes due to the lack of power, physical damage or environmental interferences. These nodes will report environment events like temperature, pressure, humidity, vehicular movement, noise levels, lighting conditions, the presence or
2
Freddy L´ opez Villafuerte et. al.
absence of certain kinds of objects, acoustic events, mechanical stress levels on attached objects, and so on. We can say that localization will act as a bridge between the virtual and physical world [2].
2
Parameters for Localization
For the different ways of estimating location information, we have to name parameters to distinguish the similarities and differences between different approaches. In this section we present the most typical parameters to classify different techniques. 2.1
Accuracy and Precision
The most important parameter for localization techniques is accuracy and precision [3]. We can define accuracy as how much the estimated position deviated from the true position is. Precision indicates how often we expect to get at least the given accuracy (e.g. 20 cm accuracy across 95% of the time). 2.2
Scalability
A location-sensing system may be developed to find tags, objects, people, assets or animals on the surface of the earth, in a city, buildings or in a single room. We can classify the location-sensing systems in a rough manner into systems which work outside and inside areas. This classification lets us identify special problems that the system will have to address such as diffraction, multipath and interference problems. Besides, the number of objects in the system plays an important role because every system has its own limit to find a number of objects per time with a given amount of infrastructure per area. An important consideration in the location-sensing systems is to select the best-fit radio frequency technology, since the increase of the objects to be localized in the network, demands for more communication and that can congest the channel if the threshold is exceeded. This characteristic is also known as responsiveness or sampling [2] and is defined as how quickly the location system outputs the location information. 2.3
Self-organization, Autonomy and System Organization
The degree of autonomy has some of the most significant consequences in the system design; this parameter is closely related to the scalability of the system. We can say that a system has high autonomy when there is little or no human intervention necessary to operate the system and the nodes act as totally independent entities. The autonomy of a system is achieved through the use of extensive and sophisticated internal processes that make their own coordination possible. The
Title Suppressed Due to Excessive Length
3
self-coordination of the network is important because it is closely related to the possibility to extend the system. We classify the autonomy and self-organization into centralized or distributed system in a rough manner, that means, the system could or may not require the help of a central entity to monitor and control the activities of the elements. 2.4
Cost
We can evaluate the cost of the location sensing system in different ways; including cost in terms of time spend for installation, money, computational effort or energy. The time cost of the system includes factors such as the length of the installation process and the needs for system administration. The capital cost of the system can be directly mapped to the money needed to set up and operate the system such as the amount of infrastructure installed, the salaries of support personnel and the maintenance of the system. The computational cost is a crucial parameter and is closely related to the location algorithm of the system. It determines the architecture of the locationsensing system which can be organized either in a centralized or a distributed manner. The centralized systems control and monitor the system functions with the help of a central engine. The systems where the location algorithm is put into each node of the network to compute its own position is called decentralized or distributed system.
3
Location techniques
Location discovery is fundamentally based on two main phases: a measurement phase that produces a set of measurements of distances, presence hops, angularly or optically to/from a set of anchor points; and a combining phase that combines the measurements to produce a final estimate of a position. We can classify the automatic location sensing into three groups: 1. Proximity-based approaches. 2. Lateration. 3. Analysis of Prior Survey Location This classification depends on the information that a node uses about a node’s neighbourhood (proximity-based approaches), the exploiting geometric properties of a given scenario (lateration), or if they are trying to analyze characteristics of the position of a node in comparison with pre-measured properties (analysis of prior survey location). Location systems may employ them individually or in combination. 3.1
Lateration
Lateration is an important concept in the localization area that computes the position of an object by measuring its distance from multiple reference positions.
4
Freddy L´ opez Villafuerte et. al.
In a plane, the position of a node can be derived with the distance measure from 3 non-collinear points. Nodes which known their exact position a priori or through the help of specialized hardware (e.g. GPS) are called anchors, landmarks, seeds, reference nodes or beacons [4]. For lateration, the extension to three dimensions requires 4 non-coplanar distances to the beacons to estimate position. Lateration uses a method known as “least squares” to estimate a particular position from a set of linearized equations in the form of AX = b. The least squares method is efficient because it minimizes possible range estimation errors accumulated along the propagation path. 3.2
Proximity
The simplest technique is to find out how “close” an object is to know its location. The object’s presence is sensed using a physical phenomenon with limited range, like a radio signal, magnetic field, or IR light. If the object is sensed and recognized within the range of the emitter then the node is able to run an algorithm to determine the approximated position of his neighbour. The proximity technique is also known as range-free approach [5]. 3.3
Analysis of prior survey location
The analysis of prior survey location techniques uses features of a scene or area observed from a particular vantage point to draw conclusions about the location of the observer or objects in the scene. In static scene analysis, observed features are looked up in a predefined dataset that maps them to object locations. In contrast, differential scene analysis tracks the difference between successive scenes to estimate location. Differences in the scenes will correspond to movements of the observer and its features in the scenes are known to be at specific positions, the observer can compute his own position relative to them [6]. Nowadays there are systems that work with a hybrid way of analysis of prior survey systems. Such systems work a priori in the covered area and extract samples of different points by (multi)lateration or proximity with the purpose of constructing a reference map of the area. Later the devices find their own location by comparison of the previous planned map with the actual readings that the devices have [7]. The major drawback of this method is that it requires substantial computational effort and some times demands a big quantity of physical memory. 3.4
Distance estimations
A system that is based on lateration, multilateration or hybrid analysis of prior survey location needs to estimate distances between the beacons and the nodes which want to discover their location. The most important approaches to estimate this distance between nodes or between nodes and landmarks are: Signal Attenuation, Time of Arrival (TOA), Time Difference of Arrival (TDOA) and Interferometry, which will be introduced in this sub-section.
Title Suppressed Due to Excessive Length
5
Signal Attenuation This method can be implemented in any system which is able to achieve radio-communication at least in an indirect way like in [8]. This method is also known as Received Signal Strength Indicator (RSSI). The general principle is to use a law of physics concerning the loss of signal strength in space. This loss of signal strength is proportional to 1/d2 where d is the distance between sender and receiver. In real scenarios the relation could be 3, 3.5, 4, etc. depending on the environment where the nodes are. The main disadvantage is that RSSI values are not constant but can heavily oscillate due to signal propagation issues such as reflection, refraction and multipaths. These cases can be found even when sender and receiver do not move. Time of Arrival This method -called also “time of flight”-, exploits the relationship between the beacon-node distance and the transmission time that a signal has to travel between sender and beacon. If the velocity of the signal is known and assuming that the sender and receptor know the time when a transmission starts, then the time of arrival of the signal is an indicative of the beacon-node distance and this distance can be computed using the propagation time by either the beacon or the node [9]. The two main disadvantages of this method are: it is necessary to have a synchronized sender and receiver. Depending on the transmission medium that is used, a high clock resolution is required to produce results of acceptable accuracy. For example, for acoustic waves, this requirement is modest (about microseconds) but for radio, a very high resolution is necessary (about nanoseconds). Time Difference of Arrival This method can be used if two transmission mediums of very different propagation speed are accessible, for example, radio waves propagating at the speed of light and ultrasound. To estimate distances, the sender must simultaneously transmit two different speed signals to the receiver. The receiver can use the arrival of the first faster signal to start measuring the time until the arrival of the second slower signal, safely ignoring the propagation time of the first one. The time registered by the receiver is proportional to the distance between sender and receiver [10]. The disadvantage to implement this approach is the necessity of two different types of receptors and senders in every node. Interferometry This approach [11] is useful if the environment where the nodes of the system work has low mobility and is free of physical obstacles. The method uses the superposition of the fields of two separate radio transmissions to create interference. Two senders and two receivers use the radio interference to obtain distance relationships between each others. The nodes create and solve the equation sets to find theirs relative positions. The solution has good accuracy (approximately 10 cm) but at the cost of relative high synchronization, Line of sight and high computational effort requirements.
6
3.5
Freddy L´ opez Villafuerte et. al.
Angles estimations
Some systems depend on Angle of Arrival (AOA) data. This method consists of the angle obtained between a reference node and the node which wants to know its position. The AOA is typically gathered using radio-chips or microphones array, which allows a listening node to determine the angular direction of a transmitting node [12]. One can also obtain AOA data from optical communication [13].This method is implemented by spatially separated microphones that hear a single transmitted signal. By analyzing the phase or time difference between the signal arrivals at different microphones, it is possible to discover the angle of arrival of the signal. Unfortunately, AOA’s hardware tends to have a big form factor and is more expensive than the distance estimation methods, since each node must have one speaker and several microphones.
4
Signal technologies
The following section gives an overview of existing technologies, their advantages and disadvantages that imply the use of the chosen technology and some examples of applications which work with the technology in study. The signal technologies that are based on infrared signals, such as the Active Badge System [14], detect the presence of an object by a blink of IR signal. This kind of system is attractive due to its inexpensive deployment, the compact factor form of the tags and the low energy consumption but it has the inconvenience that the systems are restricted to the line of sight conditions.and it has low responsiveness. The ultrasound signals are another option to estimate location; the systems based on this principle have better accuracy, some applications like [15–17] and [18] report accuracies in the order of centimetres but at the cost of work outdoors with controlled environment or a high degree of calibration and direct line of sight. Systems based on Electromagnetic signal like [19], usually work with electromagnetic fields to confirm the presence of an object. The range of this system is about 5 to 6 meters, but it has a big form factor and the infrastructure is too expensive and focused to the application. Optical signals-based systems [13] are cheap and have nodes with a compact factor form; the typical range of such systems is up to 5 meters. The main disadvantage is that the system does not work in direct sunlight and participating nodes must be in Line of sight. The most popular technology that the majority of location-sensing systems use is the radio frequency. Probably the main reason is that it does not require the line of sight, thus making a system more flexible. Depending on the type of frequency range used or its technique based transmission, radio frequency can be categorized as follows: Radio Frequency IDentification (RFID) is an automatic identification method that uses devices called RFID tags to store and remotely retrieve environment
Title Suppressed Due to Excessive Length
7
data. There are systems like [20] and [21] which use these devices to do localization. Wi.Fi. is a brand originally licensed by Wi-Fi Alliance to describe the underlying technology of Wireless Local Area Networks (WLAN ) based on the IEEE 802.11 specifications. Systems as [22, 23] and [7] use this technology to implement localization. Bluetooth is an industrial specification for Wireless Personal Area Networks (PAN) based on the IEEE802.15.1 standard. Systems like [24] and [25] use this technology to locate devices. Ultra Wide Band (UWB ) is a technology for transmitting information spread over a large bandwidth for PAN, [12] and [26] are systems that use this technology to discover the position of the nodes. Spread Spectrum for ANSI 371.1 is a technique that is used in [19] for localization, others technologies based on RF signals are Zigbee (IEEE802.15.4), DC-Electromagnetic Pulses applied in [27], Digital Television signals used in systems as [28], Commercial Radio FM Stations signals proposed by [29] and [30], wide area cellular [31] and satellite based [32]. Table 1, located at the end of the article, summarizes the different localization technologies pointing out their drawbacks and advantages.
5
Conclusion
In this survey paper we presented, different methods to implement localization and the main parameters that a system designer has to take into account to understand and value the different location-sensing systems. We also offered a classification of techniques to develop localization in systems with low power and restricted memory wireless sensor nodes, and we mentioned different available technologies to reach this goal. From this analysis we can note that the possibilities to carry out localization with different embedded sensor platforms have good perspectives, we have realized that there are more than one algorithm that can be implemented in limited sensor boards such as the exposed above. The major contribution of this paper is to identify and recognize the different ways to do location and to point out the most important characteristics of every method in order to be implemented in low power and restricted memory wireless sensor nodes.
References 1. Akyildiz, I., Su, W., Sankarasubramaniam, Y., Cayirci, E.: A survey on sensor networks. Communications Magazine, IEEE 40 (2002) 102–114 2. Estrin, D., Culler, D., Pister, K., Sukhatme, G.: Connecting the physical world with pervasive networks. Pervasive Computing, IEEE 1 (2002) 59–69 3. Hightower, J., Borriello, G.: Location systems for ubiquitous computing. Computer 34 (2001) 57–66 4. Holguer, K., Willig, A.: Protocols and Architectures for Wireless Sensor Networks. John Wiley & Sons (2005) 5. Bulusu, N., Heidemann, J., Estrin, D.: GPS-less low-cost outdoor localization for very small devices. Personal Communications, IEEE [see also IEEE Wireless Communications] 7 (2000) 28–34
8
Freddy L´ opez Villafuerte et. al.
6. Shafer, S.: Ubiquitous Computing and the EasyLiving Project. 40th Anniversary Symposium, Osaka Electro-Communications University. ht tp://www. research. microsoft. com/easyliving (2001) 7. Lenz, D.: Homepage of the ekahau project. http://www.ekahau.com/ (2007) 8. ODell, M., ODell, R., Wattenhofer, M., Wattenhofer, R.: Lost in Space Or Positioning in Sensor Networks. REALWSN, June (2005) 9. Bible, S., Zyda, M., Brutzman, D.: Using Spread Spectrum Ranging Techniques for Position Tracking in a Virtual Enviroment. Second IEEE Workshop Networked (2000) 10. Whitehouse, K.: The Design of Calamari: an Ad-hoc Localization System for Sensor Networks. University of California at Berkeley (2002) ´ Balogh, G., 11. Mar´ oti, M., V¨ olgyesi, P., D´ ora, S., Kus` y, B., N´ adas, A., L´edeczi, A., Moln´ ar, K.: Radio interferometric geolocation. Proceedings of the 3rd international conference on Embedded networked sensor systems (2005) 1–12 12. Steggles, P., Gschwind, S.: The Ubisense Smart Space Platform. Adjunct Proceedings of the Third International Conference on Pervasive Computing (2005) 13. R¨ omer, K.: The Lighthouse Location System for Smart Dust. Proceedings of the 1st international conference on Mobile systems, applications and services (2003) 15–30 14. Want, R., Hopper, A., Falc˜ ao, V., Gibbons, J.: The active badge location system. ACM Transactions on Information Systems (TOIS) 10 (1992) 91–102 15. Broxton, M., Lifton, J., Paradiso, J.: Localizing a sensor network via collaborative processing of global stimuli. Wireless Sensor Networks. Proceeedings of the Second European Workshop on (2005) 321–332 16. Priyantha, N., Chakraborty, A., Balakrishnan, H.: The Cricket location-support system. Proceedings of the 6th annual international conference on Mobile computing and networking (2000) 32–43 17. Priyantha, N.: The Cricket Indoor Location System. PhD thesis, Massachusetts Institute of Technology (2005) ´ Balogh, G., Kusy, B., N´ 18. Simon, G., Mar´ oti, M., L´edeczi, A., adas, A., Pap, G., Sallai, J., Frampton, K.: Sensor network-based countersniper system. Proceedings of the 2nd international conference on Embedded networked sensor systems (2004) 1–12 19. Wireless, S.: Home page of the wherenet project. http://http://www.wherenet.com/ (2008) 20. Amsalem, R., Aljadeff, D.: Radio-frequency identification (RFID) tag employing unique reception window and method therefor (2007) US Patent 7,295,115. 21. Chien, C.: Location position system for relay assisted tracking (2003) US Patent 6,512,478. 22. Wireless, S.: Home page of the skyhookwireless project. http://www. skyhookwireless. com/index. html (2007) 23. Intille, S., Larson, K., Beaudin, J., Nawyn, J., Tapia, E., Kaushik, P.: A living laboratory for the design and evaluation of ubiquitous computing technologies. Conference on Human Factors in Computing Systems (2005) 1941–1944 24. Papadogkonas, D., Roussos, G., Levene, M.: A Navigation Engine for Ubicomp Environments. The 4th UK-Ubinet Workshop (2006) 10–12 25. Rehrl, K., G¨ oll, N., Leitinger, S., Bruntsch, S.: Combined indoor/outdoor Smartphone navigation for public transport travellers. Location Based Services & Telecartography-Proceedings of the Symposium (2005)
Title Suppressed Due to Excessive Length
9
26. Lee, J., Scholtz, R.: Ranging in a dense multipath environment using an UWB radio link. Selected Areas in Communications, IEEE Journal on 20 (2002) 1677– 1683 27. motionstarwireless: Homepage of the motionstarwireless project. http://www.ascension-tech.com/products/motionstarwireless.php (2007) 28. Rabinowitz, M., Spilker Jr, J.: A New Positioning System Using Television Synchronization Signals. Broadcasting, IEEE Transactions on 51 (2005) 51–61 29. Krumm, J., Cermak, G., Horvitz, E.: RightSPOT: A Novel Sense of Location for a Smart Personal Object. Ubiquitous Computing: 5th International Conference, Seattle, Wa, Usa, October 12-15 212 (2003) 30. Youssef, A., Krumm, J., Miller, E., Cermak, G., Horvitz, E.: Computing location from ambient FM radio signals [commercial radio station signals]. Wireless Communications and Networking Conference, 2005 IEEE 2 (2005) 31. Laasonen, K., Raento, M., Toivonen, H., Laasonen, K., Raento, M., Toivonen, H., FI, C.: Adaptive On-Device Location Recognition. Pervasive Computing: Second International Conference, Pervasive 2004, Linz/Vienna, Austria, April 18-23 (2004) 32. Kaplan, E., Hegarty, C.: Understanding GPS: Principles and Applications Second Edition. Artech House 680 (2005) 33. Bulusu, N., Heidemann, J., Estrin, D.: Adaptive beacon placement. Proceedings of the 21st International Conference on Distributed Computing Systems (ICDCS-21) (2001) 489–498 34. Doherty, L., Pister, K., El Ghaoui, L.: Convex Position Estimation in Wireless Sensor Networks. IEEE INFOCOM, Anchorage, AK, April (2001) 35. He, T., Huang, C., Blum, B., Stankovic, J.: Range-free Localization Schemes in Large Scale Sensor Networks. proceedings of the 9th annual international conference on Mobile computing and networking (2003) 36. Mendoza, M., Silva, V., Diaz, A.: An efficient algorithm for localization in wireless sensor networks based on internal array of nodes within cells. Parallel Processing, 2005. ICPP 2005 Workshops. International Conference Workshops on (2005) 405– 412 37. Nicolescu, D., Nath, B.: Ad-Hoc Positioning Systems (APS). Proceedings of IEEE GLOBECOM 1 (2001) 25–29 38. Nagpal, R.: Organizing a Global Coordinate System from Local Information on an Amorphous Computer. Technical Report AIM-1666, MIT A.I. Laboratory (1999) 39. Nagpal, R., Shrobe, H., Bachrach, J.: Organizing a Global Coordinate System from Local Information on an Ad Hoc Sensor Network. Information Processing in Sensor Networks: Second International Workshop, Ipsn 2003, Palo Alto, Ca, Usa, April 22-23, 2003: Proceedings (2003) 40. Savvides, A., Srivastava, M., Girod, L., Estrin, D.: LOCALIZATION IN SENSOR NETWORKS. Wireless Sensor Networks (2004) 41. Savvides, A., Han, C., Strivastava, M.: Dynamic fine-grained localization in AdHoc networks of sensors. Proceedings of the 7th annual international conference on Mobile computing and networking (2001) 166–179 42. Savvides, A., Park, H., Srivastava, M.: The Bits and Flops of the N-Hop Multilateration Primitive for Node Localization Problems. Proceedings of the First International Workshop on Wireless Networks and Applications held in conjunction with Mobicom 2002 (2002) 43. Shang, Y., Ruml, W., Zhang, Y., Fromherz, M.: Localization from mere connectivity. Proceedings of the 4th ACM international symposium on Mobile ad hoc networking & computing (2003) 201–212
10
Freddy L´ opez Villafuerte et. al.
Table 1. Taxonomy of Localization Technologies. Signal Technology Infrared (IR)
Advantages Inexpensive Small Form Factor Low Energy Consumption
Drawbacks Line of Sight Requirement Low Responsiveness Sensible to Fluorescent and Sunlight Typical range of 5 m
Ultrasound
Accuracy in the order of cm Low Clock Rates System Inexpensive and Simple
Necessity of Calibration and Synchronization Temperature, humidity and pressure dependence Typical Range is between 3–15 m
Electromagnetic Field
Better than IR Signals No Line of Sight Requirement
Big Form Factor Expensive Typical Range is about 5–6 m
Optical
Cheap System Small Form Factor Low Energy Consumption
Line of Sight Requirement Sensible to Fluorescent and Sunlight Typical Range of 5 m
DC-Electromagnetic Pulses
High Accuracy in the order of mm High Responsiveness
Typical Range of up to 5 m Sensible to metallic objects Expensive
RFID
Cheap Tags Long Battery Life 3–4 m Accuracy
Very Low Computational Power Centralized System Typical Range 1–10 m
Wi.Fi. RF
Potential Reuse of Deployed Network No Line of Sight requirement
High Power Consumption Typical Range Wi.Fi. 50–180 m Sensible at Interference
Bluetooth
Low Power Consumption Low Cost Hardware Low Power Consumption Less Affected by Multipath Accuracy < 1 m
Typical Range Bluetooth 10–120 m Interference with Wi.Fi. signals UWB is limited at Bluetooth range Expensive
Low Power Consumption Cheap Transceivers Low Power Consumption No Line of Sight Requirement
Spread Spectrum for ANSI 371.1 Typical Range 10–300 m Typical Range Zigbee 10–100 m No Proper Propagation Model Exist
Digital Television
Digital Television Range is City Better Building Penetration
Expensive No licence-free band
FM Radio
Commercial FM Radio Range is City Very Cheap receptors Low Form Factor
low precision No licence-free bands
Cellular System
Typical Range is 100–150 m
GPS
Range of Satellite Based System is World Wide
Low accuracy Expensive Infrastructure It does not Work Indoors and in Urban Canyons Expensive Infrastructure Affected by Multipath and Ionosphere Propagation Delay
UWB
Spread Spectrum Zigbee
