Trends in indoor positioning
Surging smartphone market and increased use of WiFi gives new opportunities to indoor positioning technology
The demand for recognizing the user position is steadily increasing not only in the outdoor but also in the indoor. In the outdoor, the global navigation satellite system (GNSS) is available and guarantees high levels of positioning accuracy and availability. However, there is no standard method for localization even though many indoor positioning techniques have been developed.
In this short paper, market for indoor positioning is assumed to be potentially huge. There are mainly three areas this paper targets – public transportations, shopping malls and manufacturing facilities. Anyone who has been at the airport will remember having to look around everywhere to find the right terminal or find the right shop to buy something. The indoor positioning can provide more than a map to navigate inside the airport. Google has announced that they will make an effort to localize the user position in the airports . Indoor navigation needs indoor maps as well as indoor positioning technique which can guide users. In the shopping malls, the same technology for navigation can be used not only for finding shops but for finding spots in the parking lot. In large factories, tracking the position of skilled workers and equipment within the facilities and reassigning them will help achieve higher productivity. These examples well characterize the future market of indoor positioning: the indoor positioning system is expected to be useful in large areas, rather than in small areas. The manager of a shopping mall who wants to provide indoor positioning to customers, will find indoor positioning useful when shopping mall has a large number of shops. In the same sense, an administrator can expect that the factory which deals with heavy load of machines and resources will be likely to benefit from indoor positioning by tracking assets and skilled workers when building heavy machinery. As mentioned earlier, in the indoor positioning there are many techniques to localize user position. For wireless positioning measurements, the received signal strength (RSS), multicarrier phase measurement, time of flight (TOF) and angle of arrival (AOA) can be acquired . Since the number of smartphone users in the world is significantly high, and the accessibility of the indoor positioning system using smartphones will be high as well, smartphone-based localization methods are reviewed here. Smartphone sensors can be combined with the measurements aforementioned. In addition to these positioning techniques, many WiFi APs are available nowadays, so smartphone sensors and WiFi positioning techniques are also reviewed.
Following this, many RSS-based localization methods are conducted. On the other hand, TOF-based method is not suitable with WiFi system, and there are some possibilities of smartphone manufacturers adding new chip for TOFbased positioning, such as ultra-wideband (UWB) and ultrasound. Even though ultrasound based system gives better ranging accuracy than UWB, ultrasound cannot guarantee the accuracy in nonline- of-sight (NLOS) situation . This paper focuses on UWB based positioning solution. And it is possible for WiFi APs to acquire AOA measurements.
In smartphones, the sensors which include the gyroscope, accelerometer, and magnetometer are now available. And the barometer can be used to give the floor information to the users. The camera also gives the engineer chances to improve vision-based positioning technique. If wearable equipment such as Google Glass becomes popular, one can expect vision-based localization to be a good alternative for wireless positioning.
RSS-based positioning method enables the user to acquire their position without modifying the user’s smartphone. There are mainly three categories of measurement methods: distance, fingerprint, and proximity. From the pass-loss model, the distance between the user and WiFi APs could be estimated. However, estimated distance is inaccurate since the received signal suffers from channel fading. If signal power decreases more due to bad channel, estimated distance increases which causes error. If one acquires distance measurements from many APs and assigns weightings properly, the positioning accuracy will be improved. Skyhook maintains database of WiFi APs so that the user could calculate its position . Fingerprint uses radio map database premeasured. Using RSSs from many APs, the positioning accuracy is enhanced. RSSbased method using Euclidean distance or Bayesian-inference fingerprint approach is studied . Also, rank based localization method is studied. The difference between classical and rank based localization is that classical localization directly compares premeasured and RSS data directly, whereas rank based localization compares two rank vectors from radio map and RSS data . On the other hand, Nonlinear Kalman filter –based localization methods are now on being research . Kalman filter assumes that noise to be Gaussian, Unscented Kalman filter and Particle Kalman filter did not made that assumption, which makes this methods suitable to heterogeneous system . If the measurement model gets complicated, more calculation should be performed. To alleviate heavy calculation, crowdsourcing method could be a solution . However, it needs seamless connection between server and smartphone users.
Increases on the WiFi APs could cause interference between APs which causes the degeneration of positioning accuracy . And this may be the reason why many study proximity-based positioning techniques . A WiFi beacon only broadcasts its signal on the small footprint. If the signal surges over threshold, the user can estimate its position. In this method, only few calculations are needed. Its accuracy is mainly dependent on the size of the footprint. By using directional antenna, coverage can be minimized and detection probability can be enhanced. The WiFi beacon alone gives the accuracy of three over five meters. On the other hand, a beacon AP can broadcast several WiFi beacons in different direction s, which enables a system to enhance positioning accuracy and to estimate user heading . In this case, the WiFi beacon AP can be called as a multi-channel beacon AP, which gives two meter accuracy.
Another kind of localization could be based on AOA. Since changing smartphone is difficult, mobile AP could be altered instead for positioning purpose in order to perform the angle of arrival. Multiantenna is used to estimate inclination angle (or elevation angle) for localization . The simplest AOA system uses only two antennas. Using RF signals in both antennas, phase difference between two antennas could be obtained. Antenna distance and path difference (derived by phase difference) could be used to get the inclination angle of the signal. The wavelength of 2.4GHz is 12.5 centimeter; the antenna should be installed within 6 centimeter to avoid path ambiguity and (2.7 centimeter for 5.4GHz). Multiple antennas and complicated algorithm could be used to resolve integer ambiguities . The user position could be calculated from the estimated angles  or an estimated angle and height information . In either case, mobile AP could be called enhanced WiFi beacon. Nokia developed Bluetooth-based AOA solution  and Wifive Co. Ltd in Korea is now developing WiFi based AOA solution for both 2.4GHz and 5GHz.
Time of flight (TOF) uses the packet elapsed time between device A and device B. if device A and B are synchronized and device A sends a packet which contains time, device B could demodulates the packet and extract the elapsed time, which is converted later as the distance between device A and B. However, device A and B is not synchronized generally; time-bias exists in the time of flight measurements. This is called as the pseudorange, which means that the measurement is elapsed time plus unknown clock bias. If the infrastructure is synchronized such as global positioning system, the receiver clock bias can be estimated and could be cancelled. However, time synchronization between nodes needs high-cost because of the atomic clock or the wired infrastructure. Under nanosecond, synchronization is needed to acquire under meter accuracy. To alleviate this constraint, two-way ranging (TWR) is used in order to eliminate receiver clock bias. The device A sends a packet to device B, and device B returns a packet to device A and takes two TOF measurements. These measurements contain relative receiver clock bias which has the same magnitude and different sign. TWR measurement is acquired by averaging it. Furthermore, IEEE 802.15.4a standard recommends symmetric double-sided twoway ranging (SDS-TWR) which mitigates clock drifts . SDS-TWR averages two TWR measurements from device A and from device B. In order to estimate highprecision distance measurements, the peak estimation and the ranging counter should be utilized. For this purpose, PHY of ranging system should be designed in order to improve resolution of measurement . The standard specifies two kinds of PHY for ranging: chirp-spread spectrum (CSS) and impulse radio (IR) . CSS sweeps signal frequency to be higher (up-chirp) and lower (down-chirp) which consists a bit in CSS system . Nanotron produces CSSbased ranging system and chipset . IRUWB generates burst position modulation (BPM) signal which occupies low band (3.3-4.7GHz) or high band (5.9-10GHz) . Since severe regulation and signal shapes of IR-UWB makes its being a one-chip solution harder, Bespoon and Leti developed IR-UWB single chip in 2013 . These solutions are candidates for indoor location solution, and maybe augmented to the smartphone for these purposes.
Pedestrian Dead- Reckoning (PDR)
Dead reckoning (DR) is the simplified version of inertial navigation system (INS). By estimating velocity and heading, DR system updates the user position. On the other hand, pedestrian Dead-Reckoning 12 | Coordinates June 2014 detects user step and heading. Smartphone contains the accelerometer, gyroscope, and magnetometer. Step detection method depends on sensor position. If sensors are foot-mounted, zero velocity update (ZUPT) or zero angular rate update (ZARU) method could be implemented for step detection . If sensors are on a handheld smartphone, the pedometer algorithm depends on user motion . For accurate position estimation, some research focus on step length estimation , and other research study integration of PDR and other positioning method . For heading estimation, gyroscope error increases without calibration and magnetometer is vulnerable to interference from outer disturbance. For compensating heading error, vision-based method  or map-matching  method would be helpful. Some smartphones are equipped with a barometer, which enables estimate user floor .
Surging smartphone market and increased use of WiFi gives new opportunities to indoor positioning technology. The developer could use WiFi signal by measuring RSS or AOA. Some positioning techniques based on these measurement gives positioning result only in a small area. In this case, pedestrian dead-reckoning method can be used together in order to give seamless position in indoor environment. Camera and map data could enhance the accuracy of PDR. Or possibly, UWB chip could be broadly used for indoor positioning.
 “Now Google Maps Can Help You Find The Airport Check-In Desk”, TIME.com, Nov. 26, 2013.
 A. Bensky, “Wireless Positioning Technologies and Applications,” Altech House, 2008.
 Z. Sahinoglu, S. Gezici. and I. Güvenc, “Ultra-wideband Positioning Systems Theoretical Limits, Ranging Algorithms, and Protocols,” Altech House, 2008.
 J. Machaj, P. Brida and R. Piché, “Rank Based Fingerprinting Algorithm for Indoor Positioning,” Indoor Positioning and Indoor Navigation (IPIN), 2011.
 M. Kessel and M. Werner, “Automated WLAN calibration with a backtracking particle filter,” Indoor Positioning and Indoor Navigation (IPIN), 2012.
 E. N. Chartzi and A. W. Smyth, “The Unscented Kalman Filter and Particle Filter methods for Nonlinear Structural System Identification with Noncollocated Heterogeneous Sensing,” Journal of Structural control and health monitoring, January 2008.
 K. Kaji and N. Kawakuchi, “Design and Implementation of WiFi Indoor Localization baed on Gaussian mixture Model and Particle Filter,” Indoor Positioning and Indoor Navigation (IPIN), 2012.
 J. M. Lim, S. H. Yoo, K. J. Lee, and T. K. Sung, “Integration of Pedestrian DR and Beacon-AP based Location System for Indoor Navigation,” Symposium on International Global Navigation Satellite Systems Society (IGNSS), 2013
 S. H. Yoo, J. M. Lim, K. J. Lee, J.H. Kang, and T. K. Sung, “Integration of Beacon-AP and Pedestrian DR for Indoor Pedestrian Navigation,” International Symposium on Global Navigation Satellite Systems(ISGNSS), 2013.
 R. Mardiana and Z. Kawasaki, “Broadband radio interferometers utilizing sequential triggering technique for locating fast electromagnetic sources emitted from lightning,” IEEE Transactions on Instrumentation and Measurement, vol. 49, No.2, pp.376–381, April 2000.
 P. Fan and Z. R. Jing, “Parametric estimation of ultra wideband radar targets,” Journal of Systems Engineering and Electronics, Vol. 20, No 3, pp 499 – 503 , June 2009
 C. H. Lim, Y. H. Wan, B. P. Ng and C. M. S. See, “A Real-Time Indoor WiFi Localization System Utilizing Smart Antennas, ” IEEE Transactions on Consumer Electronics, VOL. 53, NO. 2, May 2007.
 J. M. Lim, K. J. Lee, J. H. Oh, S. H. Yoo, J. H. Kang and T. K. Sung, “Performance Analysis of Indoor Positioning Using Uplink AOA Measurement,” International Symposium on Global Navigation Satellite Systems(ISGNSS), 2013.
 F. Belloni, V. Ranki, A. Kainulainen, and A. Richter, “Angle-based Indoor Positioning System for Open Indoor Environments”, Proceeding of Workshop on Positioning, Navigation and Communication (WPNC), Hannover, Germany, 2009.
 IEEE Standard 802.15.4a-2007
 C. M. De Dominicis, P. Pivato and P. Ferrari, “Timestamping of IEEE 802.15.4a CSS Signals for Wireless Ranging and Time Synchronization,“ IEEE Transactions on Instrumentation and Measurement, VOL. 62, NO. 8, August 2013.
 http://www-leti.cea.fr/en/ Latest-news/BeSpoon-and-Leti- Establish-World-Record-Distance- Measurement-on-a-Single-Chip
 M. Romanovas, V. Goridko, A. Al- Jawad, M.Schwaab, M. Traechtler, L. Klingbeil,, and Y. Manoli, “A study on indoor pedestrian localization algorithms with foot-mounted sensors,” International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2012
 B. J. Shin, J. H. Lee, J. H. Kim, C. K. Kim, S. Lee, Y. T. Byun, D. H. Yun and T. J. Lee, “Motion-Awareness 3D PDR System in GPS-Denied Environment using Smartphone,” Proceedings of the 25th International Technical Meeting of The Satellite Division of the Institute of Navigation (ION GNSS), 2012.
 J.A.B. Link, P. Smith, N. Viol, K. Wehrle, “FootPath: Accurate map-based indoor navigation using smartphones,” International Conference on Indoor Positioning and Indoor Navigation (IPIN), 2011
 J. B. Kim and H. S. Jun, “Vision-Based Location Positioning using Augmented Reality for Indoor Navigation,” IEEE Transactions on Consumer Electronics, Vol. 54, No. 3, AUGUST 2008.
 M. Peter, B. Schäfer, J. A. B. Link, “Versatile Geo-referenced Maps for Indoor Navigation of Pedestrians,” Indoor Positioning and Indoor Navigation (IPIN), 2012.
 B. H. Lee, B. Harvey, T. Gallagher, “Using Barometers to Determine the Height for Indoor Positioning,” Symposium on International Global Navigation Satellite Systems Society (IGNSS), 2013