EDTTT

Information about EDTTT

Published on July 28, 2014

Author: nishtcr

Source: authorstream.com

Content

Non Line of sight problem in Mobile Location estimation: Non Line of sight problem in Mobile Location estimation NISHA MENON K ROLL NO:16 FISAT 28 July 2014 1 OUTLINE: 2 OUTLINE 28 July 2014 : Operator services Billing Network management Location based services Assistance Roadside assistance Personal or vehicle emergency Alarm management Driving Directions Tracking Tracking criminals Tracking external resources containers Monitoring Monitoring delivery process Fleet & freight tracking Personal Child Security Information Nearest service Traffic navigation help Information Directory LOCATION ESTIMATION : APPLICATION AREAS 28 July 2014 3 LOCATION ESTIMATION: DIFFERENT METHODS: 4 LOCATION ESTIMATION : DIFFERENT METHODS 28 July 2014 Locating mobile terminal in 2 dimensions : Locating mobile terminal in 2 dimensions re­quires the measurement of the LOS distance between the mobile and at least three participating BSs 28 July 2014 5 NLOS error: 6 NLOS error REFLECTION SHADOWING SCATTERING LINE-OF-SIGHT DIFFRACTION Range measurements corrupted by error 28 July 2014 NLOS error: NLOS error The major error sources in the mobile location include Gaussian measurement noise and non-line-of-sight ( NLOS) propagation error, the latter being the dominant factor. NLOS error translates the mobile’s location estimate to a biased estimate This problem has been recognized as a critical issue, possibly a “killer issue” for mobile location . In order to mitigate the effect of the measurement bias, it is necessary to develop location algorithms that are robust to the NLOS error. 28 July 2014 7 PROBLEM FORMULATION: PROBLEM FORMULATION BSs measure the time of arrival of a signal that has been sent to the mobile and then transponded back to the network. The arrival times are then converted to range measurements . ‘ m’th range is modelled as 28 July 2014 8 NLOS identification: NLOS identification It is not known which apriori range measurements (if any) contain NLOS errors. The NLOS measurements can be iden­tified , At each BS, the range measurements are first smoothed by modeling using Nth order polynomial fit 28 July 2014 9 NLOS identification: NLOS identification Smooth it as 2 hypothesis H0: NLOS is absent (only los component) H1: NLOS is present NLOS identification technique requires comparison of the standard deviation of a sample statistic to the known standard deviation of that statistic under the null hypothesis that the measurements are LOS . The stan­dard measurement noise is modeled as a zero-mean random variable Noise variance 28 July 2014 10 NLOS identification: NLOS identification . 28 July 2014 11 If the NLOS error is present along with nm (t) then the measured range have a significantly larger average devia­tion from the smoothed curve than standard deviation of a sample statistic NLOS identification: NLOS identification When the NLOS error is present, then the measured range will deviate from the smoothed curves on the average by The presence of a large standard deviation will be used to discriminate between LOS versus NLOS measurements . We reject the null hypothesis, Ho, (LOS case) for large values of 28 July 2014 12 Residual analysis rank test: Residual analysis rank test we assume that the LOS hypothesis has been rejected at one or more BSs but that there is some uncertainty about the hypothesis testing results. we can confirm our rejection of the null hypothesis by using a residual analysis rank test. Residual is defined as the difference between the measured range, rm ( ti ), and the calculated range, as follows: 28 July 2014 13 Residual analysis rank test: Residual analysis rank test 28 July 2014 14 LOS RECONSTRUCTION: LOS RECONSTRUCTION As range measurement is corrupted by the NLOS error, we employ a NLOS error correction technique for LOS reconstruction 28 July 2014 15 Conclusion: Conclusion Presented a new tracking algorithm that is capable of discriminating between LOS versus NLOS range measurements and correcting the NLOS error. Marilynn P, Wylie and Jack Holtzman , “The non-line of sight problem in mobile location estimation”, in Proc. IEEE International Conference on  Universal Personal Communication (Vol:2 ) pp. 827–831, 1996 28 July 2014 16 Reference THANK YOU: THANK YOU 28 July 2014 17

Related presentations


Other presentations created by nishtcr

High Capacity Optical networks
28. 07. 2014
0 views

High Capacity Optical networks

digital_signature
28. 07. 2014
0 views

digital_signature