Support Vector Machines for Indoor Sensor ... - Semantic Scholar
IEEE WCNC 2011 - Network
Support Vector Machines for Indoor Sensor Localization Wissam Farjow#1, Abdellah Chehri*2, Mouftah Hussein*2, Xavier Fernando#1 #1
Department or Electrical and Computer Engineering, Ryerson University 350 Victoria Street, Toronto, Ontario, M5B 2K3, Canada {[email protected]} * School Information Technology and Engineering (SITE), 800 King Edward Avenue, Ottawa, Ontario, 2
Canada, K1N 6N5. {achehri,[email protected]}
Abstract— Fingerprinting is chosen as the localization approach as fingerprinting has a higher accuracy than other approaches such as time-of-arrival or angel-of arrival. This paper introduces a positioning system based on IEEE802.15.4/ZigBee-based sensor networks. The system uses fingerprinting and employs Support Vector Machines (SVMs) to estimate node position. The system is cost-effective since it works with real deployed IEEE 802.15.4/ZigBee sensors nodes. The whole system requires minimal setup time, which makes it readily available for realworld applications. Keywords— ZigBee, Localization, Support Vector Machines.
I. INTRODUCTION With the rapid advances in wireless communication and portable devices technologies, the need for smart applications that could offer personalized services to the mobile users has attracted a lot of research interest in the past few years. This interest has led to the development of a range of services called “Location Based Services”. Location based services are aimed at making use of geospatial location information as well as user context in order to provide the end-user with useful personalised information. Several usage trends have been observed in location based services among which emergency services, information services and tracking services where the most general [1]. The accurate localization of objects and people in indoor environments has long been considered an important building block for ubiquitous computing applications [2], [3]. Most research of indoor localization systems has been based on the use of short-range signals, such as Wi-Fi [4], [5] Bluetooth [6], ultrasound [7], infrared [8], or RFID [9]. This works presents an accurate ZigBee-based indoor localization system in underground mine. Our system can be deployed on underground mines gallery equipped with a wireless sensor network. The system uses the fingerprinting technique to associate position dependent information such as the strength of the received signal with a location. The key idea that makes accurate ZigBee-based indoor localization possible is the use of two signal-strength fingerprints. We present an efficient algorithm based on support vector machines (SVM) technique for classification. The proposed algorithm is practical and scalable. The choice
978-1-61284-254-7/11/$26.00 ©2011 IEEE
of this point is motivated by the high performance of SVM. We evaluate our ZigBee-based indoor localization system in real underground mines. Experimental results show that the system achieves a good accuracy. The remainder of this paper is organized as follows. Section II contains the related work. Then, in Section III, we will present the measurements and fingerprinting data collection and simulation. We give a system description and theoretical analysis of support vector machine. Section V provides performance evaluation of the algorithm in. Finally, we will conclude with Section VI. II. RELATED WORKS For location in indoor environments, the fingerprinting technique seems the most attractive one. It gives higher localization accuracy than the other approaches such as timeof-flight especially in indoor environments with multi path effects. Also using methods such as the time-of-flight requires accurate synchronization of the clocks of both the mobile device and the access points because an error small error in time measurement can result in a relatively large error in the detected position [10]. Bahl et al. Was the first who observed that the strength of the signal from an 802.11 access point does not vary significantly in a given location. They used this observation to build RADAR [4], a system that performed localization based on which access points would be heard where, and how strongly. This was the first 802.11 fingerprinting system, and in the hallways of a small office building, fingerprints from three access points could localize a laptop within 2-3 meters of its true location. There have been improvements to RADAR’s fingerprint matching algorithm that have advanced accuracy [1], [2], [4],[11], and differentiated floors of a building with a high degree of precision [12]. In addition, commercial localization products have been built using 802.11 fingerprinting [13]. The primary aim of this paper is to design a system that would be able to locate mobile users in underground mines as well as provide them with location sensitive information in the
779
most efficient and cost effective way. The proposed system could be deployed in multiple applications. This works shows that localization system based on widearea ZigBee signal fingerprints can achieve high accuracy, and is in fact comparable to an 802.11-based implementation. The differences between our work and 802.11 fingerprinting systems are primarily due to the differences between 802.11 and ZigBee: 1. Due to good coverage of ZigBee coverage area in underground mines (up to 100 m) [14]. 2. The current trend for localization is based on IEEE 802.11standard. However the main limiting factor of localization systems for underground mines is the absence of line of sight (LOS). 3. In an underground mine, the surface and the architecture are often irregular. So, instead of employing a few relatively expensive 802.11 access points, we can use several inexpensive ad-hoc sensor nodes to estimate the position of an object. 4. In addition, it is easier to exploit the small size of ZigBee nodes compared to a IEEE 802.11 devices (see Tab. 1). Tab. 1: Comparison between different wireless nodes [15].
Cost ($ US) Size (cm3) Weight (g) Range (m)
WINS NG 2.0 node 100 5300 5400 100
iPAQ with 802.11 100 600 350 100
MICA node
Smart dust
10 40 70 30
0 is the penalty parameter of the error term. Furthermore, the kernel function K is given by:
We can change linear inseparable problem of (2) in lower dimension space into linear inseparable problem in higher dimensional space by kernel function K(x,x'). The optimization result is as follows:
4.2 Separation of Location Fingerprint The performance of indoor positioning systems depends greatly on the separation of location fingerprints. A location fingerprint corresponding to a location can be identified correctly if it is difficult to classify it (incorrectly) as another fingerprint by a pattern classifier [1]. Theoretically, a change in RSS is proportional to the logarithm of the distance between a transmitter and a receiver. Therefore, two different locations with different distances from the same transmitter node should have different average RSS values. However, in practice the RSS is a random variable that has its value fluctuating around the average value due to the dynamics in the environment. These fluctuating values can be grouped to get her as patterns of RSS at a particular location. Figure 6 shows two-dimensional plots of patterns form node A (x-axis) and node B (y-axis). The group of patterns at each location can be called the location fingerprint of that particular location. Each location fingerprint was calculated by simulation. From the plot, patterns of each location can be grouped together as a cluster. This figure indicates that RSS’s patterns can be separated by a separate cluster. Each signature is well distinct. This good indication is useful for localization mechanism.
In localization phase, the input of system is the collected RSS vector Θ'=[ RSS'A,RSS'B] of the mobile. The output of the system is the estimated location S'=(x',y') of the mobile node. We choose LS-SVM [18] to build model. LS-SVM is an improvement version of the standard SVM. It is simplecalculation, can improve convergence speed, and suitable for WSN for resource constraints. Here we chose the radial basis function (RBF) given by:
-75
RSS of Sensor B [dBm]
-80
The SVM-based localization mechanism is represented in figure 5.
-85
-90
-95
-100 -95
Loc at 10 m Loc at 20 m Loc at 30 m Loc at 40 m Loc at 60 m Loc at 70 m Loc at 80 m Measured -90
-85 -80 -75 RSS of Sensor A [dBm]
-70
-65
Fig. 6 Separation problem of seven locations in LOS scenario.
4.3 Localization Results
Fig. 5. Simulation procedure of sensor networks localization system.
The performance of the presented localization techniques will be evaluated using the CDF graph. As for the patternmatching algorithm, the SVM has been trained with the 100 location points for localization purposes. The figure 7 shows that the localization error for LOS and NLOS show very similar pattern. The proposed localization system provides almost the same localization error. This can be explaining by the high performance of fingerprinting technique. Across the two configurations, the algorithm achieves median accuracy between 0.6 and 1 meter. In
782
addition, the theoretical model for RSS used in this work gives a good result.
[9]
[10] 1
[11]
0.9 0.8
[12] Prob(x< abscissa)
0.7 0.6
[13] [14]
0.5 0.4 0.3
Measured (LOS) Simulated (LOS) Measured (NLOS) Simulated (NLOS)
0.2 0.1 0
0
0.5
1
1.5 2 Location Error (m)
2.5
3
[15] [16] [17] [18]
3.5
Fig. 7. CDF plots of the position estimation errors.
V. CONCLUSIONS This paper described the implementation of a novel algorithm, based on a set of support vector machine (SVMs) applied to ZigBee RSS fingerprinting technique in underground mining gallery. We investigated the performance of our system by measurement and simulation. The experimental and simulated results showed that the percentage of measurement points where our system could localize the mobile node with a median distance error from 0.6 to 1 meter which can be considered as good accuracy for may application. REFERENCES [1]
[2]
[3]
[4] [5]
[6]
[7]
[8]
Kamol Kaemarungsi, Prashant Krishnamurthy. “Properties of Indoor Received Signal Strength for WLAN Location Fingerprinting”. The First Annual International Conference on Mobile and Ubiquitous Systems: Networking and Services, 2004: 14-23. R. Want, A. Hopper, V. Falcao, and J. Gibbons, “The Active Badge Location System,” ACM Transactions on Information Systems, vol. 10, no. 1, pp. 91–102, Jan.1992. P. Krishnan, A. Krishnakumar, W. Ju, C. Mallows and S. Ganu, “A system for LEASE: System for location estimation assisted by stationary emitters for indoor RF wireless networks,” in IEEE Infocom, Hong Kong, March 2004. P. Bahl and V. Padmanabhan, “RADAR: An In-Building RF-Based User Location and Tracking System,” in INFOCOM, March 2000. E. Elnahrawy, X. Li and R. Martin, “The limits of localization using signal strength: a comparative study,” In Proceedings of the First IEEE International Conference on Sensor and Ad Hoc Communications and Networks, SECON 2004, pp. 406-414, Oct. 2004. L. Aalto, N. Gothlin, J. Korhonen and T. Ojala, “Bluetooth and WAP push based location-aware mobile advertising system,” in MobiSYS ‘04: Proceedings of the 2nd international conference on Mobile systems, applications, and services, pp. 49- 58. ACM Press, 2004. 29. N. Priyantha, A. Chakraborty, and H. Balakrishnan, “The Cricket Location-Support system,” in ACM International Conference on Mobile Computing and Networking (MobiCom), Boston, MA, Aug. 2000. A. Ward, A. Jones, A. Hopper, “A New Location Technique for the Active Office,” IEEE Personal Communications Magazine, Vol. 4, No. 5, October 1997. pp. 42-47.
783
J. Hightower, R. Want and G. Borriello, “SpotON: An indoor 3D location sensing technology based on RF signal strength,” Technical Report, University of Washington, February 2000. A. Bensky, S. Berliner, U. Holzkaner, E. Mirodin, et al., “Accurate distance measurement using RF techniques," Feb. 22 2005. US Patent 6,859,761. M. Youssef and A. Agrawala, “The Horus WLAN location determination system," Proceedings of the 3rd international conference on Mobile systems, applications, and services, pp. 205 -218, 2005. [31] K. Kaemarungsi and P. Krishnamurthy, “Modeling of indoor positioning systems based on location fingerprinting," INFOCOM 2004. Twenty-third Annual Joint Conference of the IEEE Computer and Communications Societies, Vol. 2, 2004. Ekahau Inc, http://www.ekahau.com, Abdellah Chehri, Hussein Mouftah, Paul Fortier, Hasnaâ Aniss. “Experimental Testing of IEEE801.15.4/ZigBee Sensor Networks in Confined Area“. CNSR'2010. pp.244~247. K. Zhao, L. Guibas, Wireless sensor networks, an information processing approach, Morgan-Kaufman 2004. https://www.silabs.com/ Vapnik, V. The Nature of Statistical Learning Theory;Berlin: Springer (1995). J.A.K, Suykens, T. Van Gestel, J. De Brabanter, B. De Moore. B, J. Vandewalle, “Least Squares Support Vector Machines”, World Scientific, Singapore, (2002).
Support Vector Machines for Indoor Sensor Localization Wissam Farjow#1, Abdellah Chehri*2, Mouftah Hussein*2, Xavier Fernando#1 #1
Department or Electrical and Computer Engineering, Ryerson University 350 Victoria Street, Toronto, Ontario, M5B 2K3, Canada {[email protected]} * School Information Technology and Engineering (SITE), 800 King Edward Avenue, Ottawa, Ontario, 2
Canada, K1N 6N5. {achehri,[email protected]}
Abstract— Fingerprinting is chosen as the localization approach as fingerprinting has a higher accuracy than other approaches such as time-of-arrival or angel-of arrival. This paper introduces a positioning system based on IEEE802.15.4/ZigBee-based sensor networks. The system uses fingerprinting and employs Support Vector Machines (SVMs) to estimate node position. The system is cost-effective since it works with real deployed IEEE 802.15.4/ZigBee sensors nodes. The whole system requires minimal setup time, which makes it readily available for realworld applications. Keywords— ZigBee, Localization, Support Vector Machines.
I. INTRODUCTION With the rapid advances in wireless communication and portable devices technologies, the need for smart applications that could offer personalized services to the mobile users has attracted a lot of research interest in the past few years. This interest has led to the development of a range of services called “Location Based Services”. Location based services are aimed at making use of geospatial location information as well as user context in order to provide the end-user with useful personalised information. Several usage trends have been observed in location based services among which emergency services, information services and tracking services where the most general [1]. The accurate localization of objects and people in indoor environments has long been considered an important building block for ubiquitous computing applications [2], [3]. Most research of indoor localization systems has been based on the use of short-range signals, such as Wi-Fi [4], [5] Bluetooth [6], ultrasound [7], infrared [8], or RFID [9]. This works presents an accurate ZigBee-based indoor localization system in underground mine. Our system can be deployed on underground mines gallery equipped with a wireless sensor network. The system uses the fingerprinting technique to associate position dependent information such as the strength of the received signal with a location. The key idea that makes accurate ZigBee-based indoor localization possible is the use of two signal-strength fingerprints. We present an efficient algorithm based on support vector machines (SVM) technique for classification. The proposed algorithm is practical and scalable. The choice
978-1-61284-254-7/11/$26.00 ©2011 IEEE
of this point is motivated by the high performance of SVM. We evaluate our ZigBee-based indoor localization system in real underground mines. Experimental results show that the system achieves a good accuracy. The remainder of this paper is organized as follows. Section II contains the related work. Then, in Section III, we will present the measurements and fingerprinting data collection and simulation. We give a system description and theoretical analysis of support vector machine. Section V provides performance evaluation of the algorithm in. Finally, we will conclude with Section VI. II. RELATED WORKS For location in indoor environments, the fingerprinting technique seems the most attractive one. It gives higher localization accuracy than the other approaches such as timeof-flight especially in indoor environments with multi path effects. Also using methods such as the time-of-flight requires accurate synchronization of the clocks of both the mobile device and the access points because an error small error in time measurement can result in a relatively large error in the detected position [10]. Bahl et al. Was the first who observed that the strength of the signal from an 802.11 access point does not vary significantly in a given location. They used this observation to build RADAR [4], a system that performed localization based on which access points would be heard where, and how strongly. This was the first 802.11 fingerprinting system, and in the hallways of a small office building, fingerprints from three access points could localize a laptop within 2-3 meters of its true location. There have been improvements to RADAR’s fingerprint matching algorithm that have advanced accuracy [1], [2], [4],[11], and differentiated floors of a building with a high degree of precision [12]. In addition, commercial localization products have been built using 802.11 fingerprinting [13]. The primary aim of this paper is to design a system that would be able to locate mobile users in underground mines as well as provide them with location sensitive information in the
779
most efficient and cost effective way. The proposed system could be deployed in multiple applications. This works shows that localization system based on widearea ZigBee signal fingerprints can achieve high accuracy, and is in fact comparable to an 802.11-based implementation. The differences between our work and 802.11 fingerprinting systems are primarily due to the differences between 802.11 and ZigBee: 1. Due to good coverage of ZigBee coverage area in underground mines (up to 100 m) [14]. 2. The current trend for localization is based on IEEE 802.11standard. However the main limiting factor of localization systems for underground mines is the absence of line of sight (LOS). 3. In an underground mine, the surface and the architecture are often irregular. So, instead of employing a few relatively expensive 802.11 access points, we can use several inexpensive ad-hoc sensor nodes to estimate the position of an object. 4. In addition, it is easier to exploit the small size of ZigBee nodes compared to a IEEE 802.11 devices (see Tab. 1). Tab. 1: Comparison between different wireless nodes [15].
Cost ($ US) Size (cm3) Weight (g) Range (m)
WINS NG 2.0 node 100 5300 5400 100
iPAQ with 802.11 100 600 350 100
MICA node
Smart dust
10 40 70 30
0 is the penalty parameter of the error term. Furthermore, the kernel function K is given by:
We can change linear inseparable problem of (2) in lower dimension space into linear inseparable problem in higher dimensional space by kernel function K(x,x'). The optimization result is as follows:
4.2 Separation of Location Fingerprint The performance of indoor positioning systems depends greatly on the separation of location fingerprints. A location fingerprint corresponding to a location can be identified correctly if it is difficult to classify it (incorrectly) as another fingerprint by a pattern classifier [1]. Theoretically, a change in RSS is proportional to the logarithm of the distance between a transmitter and a receiver. Therefore, two different locations with different distances from the same transmitter node should have different average RSS values. However, in practice the RSS is a random variable that has its value fluctuating around the average value due to the dynamics in the environment. These fluctuating values can be grouped to get her as patterns of RSS at a particular location. Figure 6 shows two-dimensional plots of patterns form node A (x-axis) and node B (y-axis). The group of patterns at each location can be called the location fingerprint of that particular location. Each location fingerprint was calculated by simulation. From the plot, patterns of each location can be grouped together as a cluster. This figure indicates that RSS’s patterns can be separated by a separate cluster. Each signature is well distinct. This good indication is useful for localization mechanism.
In localization phase, the input of system is the collected RSS vector Θ'=[ RSS'A,RSS'B] of the mobile. The output of the system is the estimated location S'=(x',y') of the mobile node. We choose LS-SVM [18] to build model. LS-SVM is an improvement version of the standard SVM. It is simplecalculation, can improve convergence speed, and suitable for WSN for resource constraints. Here we chose the radial basis function (RBF) given by:
-75
RSS of Sensor B [dBm]
-80
The SVM-based localization mechanism is represented in figure 5.
-85
-90
-95
-100 -95
Loc at 10 m Loc at 20 m Loc at 30 m Loc at 40 m Loc at 60 m Loc at 70 m Loc at 80 m Measured -90
-85 -80 -75 RSS of Sensor A [dBm]
-70
-65
Fig. 6 Separation problem of seven locations in LOS scenario.
4.3 Localization Results
Fig. 5. Simulation procedure of sensor networks localization system.
The performance of the presented localization techniques will be evaluated using the CDF graph. As for the patternmatching algorithm, the SVM has been trained with the 100 location points for localization purposes. The figure 7 shows that the localization error for LOS and NLOS show very similar pattern. The proposed localization system provides almost the same localization error. This can be explaining by the high performance of fingerprinting technique. Across the two configurations, the algorithm achieves median accuracy between 0.6 and 1 meter. In
782
addition, the theoretical model for RSS used in this work gives a good result.
[9]
[10] 1
[11]
0.9 0.8
[12] Prob(x< abscissa)
0.7 0.6
[13] [14]
0.5 0.4 0.3
Measured (LOS) Simulated (LOS) Measured (NLOS) Simulated (NLOS)
0.2 0.1 0
0
0.5
1
1.5 2 Location Error (m)
2.5
3
[15] [16] [17] [18]
3.5
Fig. 7. CDF plots of the position estimation errors.
V. CONCLUSIONS This paper described the implementation of a novel algorithm, based on a set of support vector machine (SVMs) applied to ZigBee RSS fingerprinting technique in underground mining gallery. We investigated the performance of our system by measurement and simulation. The experimental and simulated results showed that the percentage of measurement points where our system could localize the mobile node with a median distance error from 0.6 to 1 meter which can be considered as good accuracy for may application. REFERENCES [1]
[2]
[3]
[4] [5]
[6]
[7]
[8]
Kamol Kaemarungsi, Prashant Krishnamurthy. “Properties of Indoor Received Signal Strength for WLAN Location Fingerprinting”. The First Annual International Conference on Mobile and Ubiquitous Systems: Networking and Services, 2004: 14-23. R. Want, A. Hopper, V. Falcao, and J. Gibbons, “The Active Badge Location System,” ACM Transactions on Information Systems, vol. 10, no. 1, pp. 91–102, Jan.1992. P. Krishnan, A. Krishnakumar, W. Ju, C. Mallows and S. Ganu, “A system for LEASE: System for location estimation assisted by stationary emitters for indoor RF wireless networks,” in IEEE Infocom, Hong Kong, March 2004. P. Bahl and V. Padmanabhan, “RADAR: An In-Building RF-Based User Location and Tracking System,” in INFOCOM, March 2000. E. Elnahrawy, X. Li and R. Martin, “The limits of localization using signal strength: a comparative study,” In Proceedings of the First IEEE International Conference on Sensor and Ad Hoc Communications and Networks, SECON 2004, pp. 406-414, Oct. 2004. L. Aalto, N. Gothlin, J. Korhonen and T. Ojala, “Bluetooth and WAP push based location-aware mobile advertising system,” in MobiSYS ‘04: Proceedings of the 2nd international conference on Mobile systems, applications, and services, pp. 49- 58. ACM Press, 2004. 29. N. Priyantha, A. Chakraborty, and H. Balakrishnan, “The Cricket Location-Support system,” in ACM International Conference on Mobile Computing and Networking (MobiCom), Boston, MA, Aug. 2000. A. Ward, A. Jones, A. Hopper, “A New Location Technique for the Active Office,” IEEE Personal Communications Magazine, Vol. 4, No. 5, October 1997. pp. 42-47.
783
J. Hightower, R. Want and G. Borriello, “SpotON: An indoor 3D location sensing technology based on RF signal strength,” Technical Report, University of Washington, February 2000. A. Bensky, S. Berliner, U. Holzkaner, E. Mirodin, et al., “Accurate distance measurement using RF techniques," Feb. 22 2005. US Patent 6,859,761. M. Youssef and A. Agrawala, “The Horus WLAN location determination system," Proceedings of the 3rd international conference on Mobile systems, applications, and services, pp. 205 -218, 2005. [31] K. Kaemarungsi and P. Krishnamurthy, “Modeling of indoor positioning systems based on location fingerprinting," INFOCOM 2004. Twenty-third Annual Joint Conference of the IEEE Computer and Communications Societies, Vol. 2, 2004. Ekahau Inc, http://www.ekahau.com, Abdellah Chehri, Hussein Mouftah, Paul Fortier, Hasnaâ Aniss. “Experimental Testing of IEEE801.15.4/ZigBee Sensor Networks in Confined Area“. CNSR'2010. pp.244~247. K. Zhao, L. Guibas, Wireless sensor networks, an information processing approach, Morgan-Kaufman 2004. https://www.silabs.com/ Vapnik, V. The Nature of Statistical Learning Theory;Berlin: Springer (1995). J.A.K, Suykens, T. Van Gestel, J. De Brabanter, B. De Moore. B, J. Vandewalle, “Least Squares Support Vector Machines”, World Scientific, Singapore, (2002).
