 Research
 Open Access
Interval type 2 fuzzy localization for wireless sensor networks
 Noura Baccar^{1}Email author and
 Ridha Bouallegue^{2}
https://doi.org/10.1186/s1363401603404
© Baccar and Bouallegue. 2016
 Received: 10 August 2015
 Accepted: 23 March 2016
 Published: 2 April 2016
Abstract
Indoor localization in wireless sensor networks (WSN) is a challenging process. This paper proposes a new approach to solve the localization problematic. A fuzzy linguistic localization scheme is proposed. Based on interval type 2 fuzzy logic (IT2FL), a signal processing of the radio signal strength indicator (RSSI) minimizes the uncertainty in RSSI measurements from anchors caused by the indoor obstacles. The fuzzy system subdivides the map on fuzzy sets described by a new fuzzy location indicator (FLI). Fluctuations on RSS fingerprints are then reduced thanks to the IT2FL in the input side and the FLI in the output side. Experimentations were done in the Cynapsys indoor environment on a WSN test bed. The experimental results prove higher success rate in position estimations thanks to the FLI concept and the superiority of interval type 2 fuzzy logic to handle signal fluctuations.
Keywords
 Interval type 2 fuzzy logic
 Signal processing
 Wireless sensor network
 Localization
1 Introduction
Do we really need x, y, and z coordination for indoor localization? When we are subjected to human localization process, we refer generally to linguistic localization (near to the desk, next to the window in front of the TV…). Thus, a fuzzification of this problematic will change our angle of view and enlarge our perception of localization methods from Euclidean geometrical equations and signal propagation models to more opened intelligent and pervasive computing.
The availability and diversity of wireless communication (wireless area networks (WAN), wireless sensor networks (WSN)…) give researchers a huge amount of creative areas. Those available transmitted data and indicators may be exploited in applications to facilitate human being life. Smart buildings, smart homes, and ubiquitous cities are the trends of leading projects in pioneer companies. Hence, the progress in WSN deployment for “smart” purposes, besides the implication of those huge technological companies in this field, give a big motivation to innovate in various communication techniques. Mobility and localization are indeed two constraining factors relative to WSN design problematic. Many emerging contextaware applications are stand on locationbased services (LBS).
Since that, geolocalization in WSN has been the subject of many researches. For outdoor as well as indoor environment, the use of computational intelligence in localization techniques is not a new invented methodology. However, the specific nature of the indoor environment (shadowing, reflection, path loss…) originates depth investigation using different optimization techniques. Type 1 fuzzy logic (T1FL) is one of those techniques for geometrical localization and as a clusteringbased methodology. Nevertheless, there has not been any attempt to investigate the usage of interval type 2 fuzzy logic (IT2FL) in indoor geolocalization. But, it was proved that the use of IT2FL in complex realword applications presenting a high level of uncertainty in measurements performs better [1–3]. In control theory, some industrial application results were presented in [4], and the type 2 fuzzy logic controller (FLC) was applied to three domains: industrial control, mobile robot control, and ambient intelligent environment control. The author proved that type 2 FLC for each application provides smooth responses outperforming always the type 1 counterparts. This is due to the powerful paradigm of type 2 FLC to handle the high level of uncertainties present in realworld environments.
In [5], a datadriven IT2 fuzzy logic modeling framework is presented and very good computational efficiency was demonstrated through real industrial case study posing particular challenges in terms of data uncertainty comparing to type 1 fuzzy logic. The superiority of IT2FL to type 1 in handling measurement uncertainties in realworld applications was also proved in [6] and [7].
Since the localization problematic that we discuss here is based on radio signal strength indicator which is submitted to a high level of fluctuations and uncertainty in indoor environments and Interval Type 2 Fuzzy Localization System (IT2FLS) was proved to give better results dealing with data uncertainty, the use of IT2FLS may give similar results on the radio frequency (RF)based localization.
In this paper, a new approach on fuzzy geolocalization is proposed. Based on a linguistic concept, the expert builds an adaptive fuzzy model to the target environment. In a first learning stage, he defines the distribution of anchor nodes in a manner to cover all target space. Then, he ranges the radio signal strength indicator (RSSI) using linguistic fuzzy descriptors {low, medium, or high}. Because of the instability in the indoor RSSI measurements, an IT2FL processing is programmed. On the other hand, the expert clustered the target map on fuzzy sets using a new fuzzy location indicator (FLI). Thus, for each FLI, the expert takes the RSS fingerprints and proceeds to rule base building. Through semantic relations, the geometrical map dispositions are fuzzified to an “ifthen” linguistic description. In the online stage, signals are submitted to IT2 fuzzification and then aggregated using the inference engine of the fuzzy localization system (FLS). The FLI is defuzzified to a crisp value describing the location zone in the map. Experiments in the Cynapsys indoor environment have proved the effectiveness of this approach.
This paper is organized as follows. Section 2 will present the background of localization algorithm based on fuzzy logic. Section 3 will detail the proposed approach concepts. In Section 4, the experimentation process and the results are discussed. Finally, the paper is summarized by a conclusion and perspectives.
2 Background
Two main technological choices are basic for the design of localization systems: the localization technology and the positioning technique. Firstly, locationbased systems are generally RFbased technologies. Thanks to the speedy progress of nanotechnology [8] (the ease to reach receivers and sensors) that heavily involves shortrange communications notably WiFi, Bluetooth, and ZigBee, localization applications become available and some are based on hybrid systems, from robotic guiding [9] to locationbased services. In their survey [10], Liu et al. present existent indoor applications in the market and their different performance criteria. They concluded that fingerprinting schemes are better on indoor open areas.
Secondly, the localization techniques can be classified in three categories. The first one consists of deterministic techniques, classified as geometrical methods. They are rangebased and estimate the target coordination through multilateration, triangulation, angulation, angle of arrival (AoA), and time of arrival (TOA) needing most of the time specific hardware. In their work, Yan et al. [11] presented a fuzzybased geometrical probabilistic method to deal with nonlightofsight (NLOS) conditions. Although it presents good results, their algorithm needs complex calculations and depends on the known and precision of anchor coordination.
On the other side, a big number of research works consider the probabilistic approaches [10] like Bayesian algorithms [12] and a third localization process is based on machine learning approaches [13–15], using SVM [16] and neural networkbased algorithms [17]. In this category, Knearest neighbor (KNN) classification was deeply investigated in fingerprinting algorithms [14, 18, 19]. It shows promoting results in terms of offering adequate estimation accuracy; however, a big number of anchors are required to reach this accuracy.
Fuzzy logic was exploited in two manners: geometric fuzzification concept [20] and rulebased fuzzification concept. Wang et al. [20] demonstrate that the fuzzy geometric approach outperforms the traditional least squares approach. However, this approach is costly while it requires velocity and azimuth angle measurements. Besides, the system is only adaptable to linear trajectories and not for the various kinds of fuzzy observer trajectories. Fan et al. [21] suggest the use of fuzzy logic in a recursive least squares filter. Although it proposes a different methodology to process noise unlike statistical models, it uses the classical coordinatebased localization technique in a way raising the filtering processes leading to the increase of computational complexity.
GarciaValverde et al. in [22] and [23] have worked on a mobile application based on fuzzy logic. They build an adaptive rulebased model to the system. Through a T1 fuzzification in RSS classification, the used technique is able to automatically learn offline and online to adapt in order to deal with the environmental changes. It reaches 82.22 % of success accuracy. But to handle RSS fluctuations, they used two alternatives: a heuristic preprocessing algorithm and a responsive universal based on trimlike operation to remove peaks and drops and on a modified standard deviationbased technique applied to the last received RSSI values for every access point. On the other side, paper [24] investigates the RSSbased rangefree fuzzy ring method. This approach proves good performance face to radio propagation irregularity. However, it is computational intensive and difficult in experimental deployments while it depends on the propagation model parameter estimation.
In [25], the author uses T1FL for RSS clustering and the FLS creates linear equations to estimate the location zone. This system provides 95 % accuracy in positioning, whereas it has big granularity in localization (zones and not rooms) which are relatively away from each other and without providing results in case of adjacent rooms.
Furthermore, IT2FL was not used for localization. It is used generally in WSN for clustering sensed data. Although Liang and Wang in [26] presented a methodology to simulate uncertainty on RSSi and to cluster measurements, it was not exploited on localization. In literature, no real experiments on indoor mobile application based on type 2 fuzzy logic were found and this is the main contribution of this paper.
Our anchorbased proposed approach provides high granularity in location definition. It uses no preprocessing algorithm but the IT2 in the fuzzification phase which will handle RSS fluctuations. The following section will give all details of the proposed approach.
3 Proposed approach
In each FLI, RSS measurements are collected from anchor nodes in the data base (DB). Those measurements are processed using interval type 2 fuzzy logic algorithm to minimize instability of this indicator through its footprint of uncertainty (FOU) property. Based on the created DB, the expert extracts the linguistic rules to form the rule base of the fuzzy process.
3.1 Fuzzy interval type 2 input processing
An intelligent localization algorithm is proposed based on “interval type 2 fuzzy logic” for input processing. Type 2 FLS is generally used when the circumstances are too unknown to determine exact membership grade like when the training data is affected by noise. In a big number of control and clustering applications, higher accuracy has been proved using the IT2 fuzzy logic [27]. Satvir et al. [27] proved the viability of interval type 2 over type 1 FLSs through implemented systems in real environment. IT2 handled the presented noise by its uncertainty modeling property.
In their work [28], Aladi et al. demonstrated the relationship between the FOU size and the amount of uncertainty and noise in a given environmental setting. Thus, considering the target space as fuzzy sets will incorporate fluctuations in RSS measurements. An RSS vector in a zone will not be as specific as it was saved in the learning phase. The fuzzification process will limit this specificity and takes into consideration the instability of signal propagation in the indoor environment.
Two approaches exist to design an IT2FLS: the first approach is partially dependent and based on a type 1 FLS design and then a translation to the IT2FLS. Thus, a faster comparative study between T1 and the IT2 could be easily done. The second approach relies on a direct design of predefined IT2FLS parameters and thus avoids the effect of translation from T1 which may not give the best results.
As in this work, we intend to compare the results of the T1 and IT2 in localization error, and to prove the superiority of IT2 upon T1, the first approach is used. Thus, we will preserve the basic structure (the number of membership functions and the rule base).
The FOU is not uniform in the FS’s region. With an assumption that the noise/uncertainty is uniform, a FOU construction method gave rise to an equal amount of uncertainty in the memberships.
Gaussian membership functions are considered in both T1 and IT2FLS.
3.2 Output processing: the FLI
Through a linearization process of the 2D plan, the tagged environment “E” will be hierarchized in N fuzzy sets, Z fuzzy subsets (FLI).
The FLI specification will be equivalent to the FSs of Rssi vector in each room.
The variation in each FLI will specify the interval of each membership function. For FLI = 1, an interval of the lowest measured Rssi for FLI = 1 and the higher measured one for the same FLI will be determined.
3.3 FLS
4 Experimental results and discussion
4.1 Experimental test bed
Let us consider the plan of Cynapsys as our test bed. Three STM32W108 boards are deployed as anchors. Figure 9 shows there dispositions in a way their range covers all the target zones.
The mobile node MB954 is attached to an i5 Pc through a USB cable. The expert takes five fingerprints in each indicated zone along the target space: {Open_Space, Pythagore_meeting_room, Pythagore_corridor, Reception, Descartes_meeting_room, Descartes_corridor, RD_room}.
The localization process is composed of two main phases. The first one is the “learning phase.” In this stage, the expert saves the fingerprints relative to each room. In the second stage, the system proceeds to a fuzzy localization process.
4.2 Software developments
4.2.1 Network creation
The implementation of the localization platform was started by connecting the STM32W108 boards and building the network. In the first place, the choice of network topology was made based on the application needs. It is necessary to keep in obvious fact that the mobile board must be able to communicate directly with anchors so that we obtain the value of the RSSI between these two nodes and not with regard to another node which has broadcast the message. Based on the optimized MAC library IEEE 802.15.4, three possible topologies for the network are possible. The first one is star topology. In this type of network, all nodes are directly connected to the coordinator. Thus, there is no direct communication if we place more than one mobile. The second topology is in a tree where nodes which are in various connections of the tree are not in direct contact what does not suit our application. The final proposed topology is meshed where nodes are in direct contact. This topology suits our case.
Hence, the embedded program is charged of RSSI collection and sending. The functions realized by the embedded system are the following ones: firstly, to establish a wireless network, then connecting cards on the same network; secondly, to assure a stable communication between them. In addition to making requests to collect RSSI from various anchors, it sends the RSSI vector collected and realizes commands reading via USB to manage card behavior.
The “Simple MAC” library is used for the program. One of its functions is the callback function named “ST_RadioReceiveIsrCallback” which is going to serve us to get back the value of the RSSI. The callback function types are generally functions of interruption which runs automatically in case of an event. In our case, the “ST_RadioReceiveIsrCallback” function runs in the reception of a message. It allows getting back the package in question, the time of arrival, if it contains errors, and the RSSI. Thanks to this function, the recovery of the RSSI of every message is possible; thus, it is enough to send any message to be able to know the RSSI with the transmitter.
4.2.2 Input fuzzification
4.2.3 Output FLI definition

"MF1 = ‘Open_Space’:‘Triangular’,[0 1.5 3]"

"MF2 = ‘Pythagore_room’:‘Triangular’,[2.5 3.5 5]"

"MF3 = ‘Pythagore_corridor’:‘Triangular’,[4.5 6 8]"

"MF4 = ‘Reception’:‘Triangular’,[7 9 11]"

"MF5 = ‘Descartes_room’:‘Triangular’,[10 12 14]"

"MF6 = ‘Descartes_corridor’:‘Triangular’,[13 15 17]"

"MF7 = ‘RD_room’:‘Triangular’,[16 18 20]"
4.2.4 Rule base creation
 1.
IF (ZRSS1 IS Low) (ZRSS2 IS Low) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 2.
IF (ZRSS1 IS Low) (ZRSS2 IS Low) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 3.
IF (ZRSS1 IS Low) (ZRSS2 IS Low) (ZRSS3 IS High)THEN (FLI IS Open_Space)(1)
 4.
IF (ZRSS1 IS Low) (ZRSS2 IS Medium) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 5.
IF (ZRSS1 IS Low) (ZRSS2 IS High) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 6.
IF (ZRSS1 IS Medium) (ZRSS2 IS Low) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 7.
IF (ZRSS1 IS Medium) (ZRSS2 IS Low) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 8.
IF (ZRSS1 IS Medium) (ZRSS2 IS Low) (ZRSS3 IS High)THEN (FLI IS Open_Space)(1)
 9.
IF (ZRSS1 IS Medium) (ZRSS2 IS Medium) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 10.
IF (ZRSS1 IS Medium) (ZRSS2 IS Medium) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 11.
IF (ZRSS1 IS Low) (ZRSS2 IS Medium) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 12.
IF (ZRSS1 IS Low) (ZRSS2 IS Medium) (ZRSS3 IS High)THEN (FLI IS Open_Space)(1)
 13.
IF (ZRSS1 IS Low) (ZRSS2 IS High) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 14.
IF (ZRSS1 IS Medium) (ZRSS2 IS High) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 15.
IF (ZRSS1 IS Medium) (ZRSS2 IS High) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 16.
IF (ZRSS1 IS High) (ZRSS2 IS Low) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 17.
IF (ZRSS1 IS High) (ZRSS2 IS Low) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 18.
IF (ZRSS1 IS High) (ZRSS2 IS Low) (ZRSS3 IS High)THEN (FLI IS Open_Space)(1)
 19.
IF (ZRSS1 IS High) (ZRSS2 IS Medium) (ZRSS3 IS Low)THEN (FLI IS Open_Space)(1)
 20.
IF (ZRSS1 IS High) (ZRSS2 IS Medium) (ZRSS3 IS Medium)THEN (FLI IS Open_Space)(1)
 21.
IF (ZRSS1 IS High) (ZRSS2 IS Medium) (ZRSS3 IS High)THEN (FLI IS Open_Space)(1)
4.2.5 Experimental platform
4.3 Experimental scenarios
Then, relying on the RSSI measurements from the ZigBee anchors Z_{1}, Z_{2}, and Z_{3}, the system calculates the FLI, and using the algorithm below, the localization process is evaluated.
4.4 Results and discussion
While this work is based on a linguistic localization, the performance of the design is evaluated through the success and failure rate of the estimated position in the zone level and in the room level.
Experimental results for FLI = 20
Scenarios  Total positions  True estimated position  True estimated room  Success rate in the zone level (%)  Success rate in the room level (%) 

Fingerprinting trajectory  120  105  107  87.5  89.16 
Red trajectory  65  57  57  87.69  87.69 
Orange trajectory  18  16  16  88.88  88.88 
Purple trajectory  30  24  25  80  83.33 
Blue trajectory  30  25  25  83.33  83.33 
Experimental results for FLI = 130
Scenarios  Total positions  True estimated position  True estimated room  Success rate in the zone level (%)  Success rate in the room level (%) 

Fingerprinting trajectory  120  98  110  81.66  91.66 
Red trajectory  65  55  59  84.61  90.76 
Orange trajectory  18  16  16  88.88  88.88 
Purple trajectory  30  24  28  80  93.33 
Blue trajectory  30  22  25  73.33  83.33 
Further experiments were conducted on the R&D room (8 m × 5.5 m) to calculate the average localization error (The localization error is measured as the Euclidean distance between the actual and the estimated locations). For three anchor boards (STM32W108), 200 fingerprints were taken through 40 FLI in the target zone. Thirtyfive locations were considered for evaluation. For the sake of consistency and completeness, we use our gathered data to evaluate our proposed IT2FL algorithm and two nonfuzzy algorithms: the KNNbased localization method and the lateration algorithm.
The KNN method is based on comparative searches of the profiled fingerprints to choose the K closest profiled samples in terms of minimizing the RSS discordance between the query RSS sample and the profiled ones. The weighted coordinates of these K samples generate the estimated location.
Comparative experimental results of the average localization error between fuzzy and nonfuzzy algorithms
Localization algorithm  Average localization error (m) 

IT2FLS  0.8 
Knearest neighbor (KNN)  1.1 
Lateration  1.9 
5 Conclusions
This paper proposed a linguistic fuzzy modeling focused on interpretability for localization of mobile nodes in wireless sensor networks. The uncertainty in the linguistic localization system was processed in two ways: in the first place, an interval type 2 fuzzification was proposed to handle RSSI fluctuations. Secondly, a fuzzy location indicator (FLI) was considered to handle geometric repartitions of fuzzy fingerprints. Experimentations have proved a high success rate either for the zone level or the room level. Besides, the superiority of IT2FL to T1FL to minimize RSSI uncertainties has been proved.
In the future work, we intend to work on the automatic generation of the rule base through a neurofuzzy algorithm (GARIC), in addition to automating the FLI recording through a preprogrammed drone.
Declarations
Acknowledgements
These works of research and innovation are made within a MOBIDOC thesis, financed by the European Union (EU) within the framework of the PASRI program, and partially supported by Cynapsys IT Enterprise. We give our thanks to Mootez Jridi and Meher Houidi for their assistances.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
Authors’ Affiliations
References
 R Martinez, O Castillo, LT Aguilar, Optimization of interval type2 fuzzy logic controllers for a perturbed autonomous wheeled mobile robot using genetic algorithms. Inf Sci 179(13), 2158–2174 (2009)View ArticleMATHGoogle Scholar
 CH Hsu, CF Juang, Evolutionary robot wallfollowing control using type2 fuzzy controller with speciesDEactivated continuous ACO. IEEE Trans. Fuzzy Syst. 21(1), 100–112 (2013)View ArticleGoogle Scholar
 R John, P Innocent, M Barnes, Type 2 fuzzy sets and neurofuzzy clustering of radiographic tibia images. Proceedings of the Sixth IEEE International Conference in Fuzzy Systems 3, 1375–1380 (1997)View ArticleMATHGoogle Scholar
 H Hagras, Type2 FLCs: a new generation of fuzzy controllers. IEEE Comput. Intell. Mag. 2(1), 30–43 (2007)View ArticleGoogle Scholar
 A Rubio Solis, G Panoutsos, Interval type2 radial basis function neural network: a modelling framework. IEEE Trans. Fuzzy. Syst. 23(2), 457–473 (2015)View ArticleGoogle Scholar
 R Sepúlveda, O Castillo, P Melin, A Diaz, O Montiel, Handling Uncertainty in Controllers Using Type2 Fuzzy Logic. Proceeding of the 2005 IEEE International Conference on Fuzzy Systems,USA, 2005, pp. 248–253Google Scholar
 J Mendel, Type2 fuzzy sets and systems: an overview. IEEE Comput. Intell. Mag. 2(2), 20–29 (2007)View ArticleGoogle Scholar
 ABI . Research, “Indoor location technologies,” US, AN1331 https://www.abiresearch.com/marketresearch/product/1016891indoorlocationtechnologyoems/. (2014)
 P Steinhaus, M Strand, R Dillmann, Autonomous robot navigation in humancentered environments based on 3D data fusion. EURASIP J. Adv. Signal Process. 2007(86831), 1–10 (2006)MATHGoogle Scholar
 H Liu, H Darabi, P Banerjee, J Liu, Survey of wireless indoor positioning techniques and systems. IEEE T. Syst. Man CY. C. 37(6), 1067–1080 (2007)View ArticleGoogle Scholar
 J Yan, K Yu, L Wu, Fuzzy modeling, maximum likelihood estimation, and Kalman filtering for target tracking in NLOS scenarios. EURASIP J. Adv. Signal Process. 2014, pp. 116 (2014)Google Scholar
 M.R. Morelande, B. Moran, M. Brazil, “Bayesian node localisation in wireless sensor networks,” in Acoustics, Speech and Signal Processing. ICASSP 2008. IEEE International Conference on, Las Vegas pp. 25452548 , (2008)Google Scholar
 H. Ahmadi and R. Bouallegue, “Comparative study of learningbased localization algorithms for Wireless Sensor Networks: Support Vector regression, Neural Network and Naïve Bayes,” Wireless Communications and Mobile Computing Conference (IWCMC) proceedings, Dubrovnik, pp. 1554–1558, 2015.Google Scholar
 IT Haque, C Assi, Profilingbased indoor localization schemes. IEEE Syst. J. 9(1), 76–85 (2015)View ArticleGoogle Scholar
 MA Alsheikh, S Lin, D Niyato, HP Tan, Machine learning in wireless sensor networks: algorithms, strategies, and applications. IEEE Commun. Surv. Tutorials 16(4), 1996–2018 (2014)View ArticleGoogle Scholar
 A Ukil, Support Vector Machine, Chapter 4, Intelligent Systems and Signal Processing in Power Engineering (SpringerVerlag, Berlin, 2007)Google Scholar
 L Gogolak, S Pletl, D Kukolj, Neural networkbased indoor localization in WSN environments. Acta. Polytechnica. Hungarica. 10(6), 221–235 (2013)Google Scholar
 P. Bahl and V. N. Padmanabhan, “RADAR: an inbuilding RFbased user location and tracking system,” in INFOCOM 2000. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE pp 775–784 vol.2, Tel Aviv (2000). http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=832252&newsearch=true&queryText=RADAR:%20an%20inbuilding%20RFbased% 20user%20location%20and%20tracking%20system
 T. Oktem, D. Slock, “Pairwise error probability analysis for power delay profile fingerprinting based localization, pp. 1–5.,” in Vehicular Technology Conference (VTC Spring), 2011 IEEE 73rd, Yokohama (2011). http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arnumber=5956780
 R. Wang, W. Cao, W. Wan, Location discovery based on fuzzy geometry in passive sensor networks. Int Journal of Digital Multimedia Broadcasting. , vol. 2011, Article ID 851951, 6 pages. doi:10.1155/2011/851951 Hindawi Pub. Co, (2011)Google Scholar
 E Fan, Wx Xie, Zx Liu, Maneuvering target tracking using fuzzy logicbased recursive least squares filter. EURASIP J. Adv. Signal Process. 53(53), 1–16 (2014)Google Scholar
 T GarciaValverde, A GarciaSola, A GomezSkarmeta, J Botia, H Hagras, J Dooley, V Callaghan, "An adaptive learning fuzzy logic system for indoor localisation using WiFi in ambient intelligent environments". Proc. FUZZIEEE 1(8), Brisbane, pp 10–15 (2012)Google Scholar
 T GarciaValverde, A GarciaSola, H Hagras, J Dooley, V Callaghan, J Botia, A fuzzy logicbased system for indoor localization using WiFi in ambient intelligent environments. IEEE Trans. Fuzzy Systems 21(4), 702–718 (2013)View ArticleGoogle Scholar
 A Velimirovic, G Lj, M Djordjevic, M Velimirovic, M Jovanovic, A fuzzy set based approach to rangefree localization in wireless sensor networks. Ser.:Elec.Energ. 23(2), 227–244 (2010)Google Scholar
 A. Salazar, A. L and G. Licea, “Estimating indoor zonelevel location using WiFi RSSI fingerprinting based on fuzzy inference system,” ICMEAE, Conf proceedings, Morelos pp. 19–22, 2013.Google Scholar
 Q. Liang and L. Wang, "Sensed Signal Strength Forecasting for Wireless Sensors Using Interval Type2 Fuzzy Logic System," in The 14th IEEE International Conference on Fuzzy Systems (FUZZ '05) pp. 2530., 2005. Google Scholar
 S. Satvir, S. G. Inderjeet, S. Sarabjeet and D. Gaurav, “Application of type2 fuzzy logic—a review,” International Conference on Communication, Computing & Systems(ICCCS) proceedings, Punjab pp. 259–263, 2014.Google Scholar
 J. Aladi, C. Wagner, J. Garibaldi, “Type1 or interval type2 fuzzy logic systems—on the relationship of the amount of uncertainty and FOU size,” Fuzzy Systems (FUZZIEEE), 2014 IEEE International Conference on, Beijing pp. 2360–2367, 2014.Google Scholar
 A. Sadeghian, J. Mendel, H. Tahayori, “Advances in type2 fuzzy sets: theory and applications”, Springer. Vol 30 (2013)Google Scholar
 JM Mendel, Uncertain RuleBased Fuzzy Logic Systems Introduction and New Directions (Prentice Hall, Indiana ᅟ, 2001). http://www.informit.com/store/uncertainrulebasedfuzzylogicsystemsintroduction9780130409690.
 K Langendoen, N Reijers, Distributed localization in wireless sensor networks: a quantitative comparison. J. Comput. Netw. 43(3), 499–518 (2003). http://dx.doi.org/10.1016/S13891286(03)003566 View ArticleMATHGoogle Scholar