Error correction of predicted signal levels in mobile communications: Master's thesis

Wenche Backman

Research output: ThesisMaster's thesisTheses

Abstract

The aim of this work is to combine measured and predicted signal levels in order to improve the accuracy of a propagation model. In addition to inserting measured signal levels in the predicted coverage area matrix, new values og signal level are interpolated around the measured ones and inserted in the matrix. Three interpolation techniques have been evaluated: Moving Least Squares (MLS), Delaunay Triangulation and Averaging. The first two methods interpolate signal level values over a wide area with samples from only a few measurement points, while averaging requires closely spaced, evenly scattered points. The purpose of this work has been to build up a database of signal levels for mobile terminal location. In addition, the improved coverage area predictions can be used for network planning purposes. The standard deviation is of primal interest when the accuracy of the interpolated signal levels is assessed because a small standard deviation proves the consensus between prediction and measurementin signal level changes. For the propagation model used in this work, the mean of the standard deviation of the error for seven test cases is 5.6 dB, compared to 6.1 dB with the MLS method and 5.1 dB with Triangulation. In cases where there were no shadowing obstacles present and there were smooth changes in the signal level, the MLS method gave better results while Triangulation performed better in areas with abrupt changes in the signal level. Hence, MLS should be used for error correction in rural and suburban areas while Triangulation performs better in urban areas, metropolises and probably also indoors. The test area is considered urban. A comparison between all three methods revealed that in cases of closely spaced uniformly distributed points Averaging should be used because it gives the smallest mean value of the standard deviaion (3.0 dB). Furthermore, with the standard deviation of the error ranging from 2 to 11 dB interpolation techniques can be used for signal level prediction purposes themselves.
Original languageEnglish
QualificationMaster Degree
Awarding Institution
  • Helsinki University of Technology
Place of PublicationEspoo
Publisher
Publication statusPublished - 2003
MoE publication typeG2 Master's thesis, polytechnic Master's thesis

Fingerprint

Error correction
Triangulation
Communication
Interpolation
Planning

Keywords

  • signal level interpolation
  • signal level prediction
  • propagation model
  • coverage area
  • mobile communications

Cite this

Backman, W. (2003). Error correction of predicted signal levels in mobile communications: Master's thesis. Espoo: Helsinki University of Technology.
Backman, Wenche. / Error correction of predicted signal levels in mobile communications : Master's thesis. Espoo : Helsinki University of Technology, 2003. 66 p.
@phdthesis{a87fb5e0493748938c0ba48138f7961b,
title = "Error correction of predicted signal levels in mobile communications: Master's thesis",
abstract = "The aim of this work is to combine measured and predicted signal levels in order to improve the accuracy of a propagation model. In addition to inserting measured signal levels in the predicted coverage area matrix, new values og signal level are interpolated around the measured ones and inserted in the matrix. Three interpolation techniques have been evaluated: Moving Least Squares (MLS), Delaunay Triangulation and Averaging. The first two methods interpolate signal level values over a wide area with samples from only a few measurement points, while averaging requires closely spaced, evenly scattered points. The purpose of this work has been to build up a database of signal levels for mobile terminal location. In addition, the improved coverage area predictions can be used for network planning purposes. The standard deviation is of primal interest when the accuracy of the interpolated signal levels is assessed because a small standard deviation proves the consensus between prediction and measurementin signal level changes. For the propagation model used in this work, the mean of the standard deviation of the error for seven test cases is 5.6 dB, compared to 6.1 dB with the MLS method and 5.1 dB with Triangulation. In cases where there were no shadowing obstacles present and there were smooth changes in the signal level, the MLS method gave better results while Triangulation performed better in areas with abrupt changes in the signal level. Hence, MLS should be used for error correction in rural and suburban areas while Triangulation performs better in urban areas, metropolises and probably also indoors. The test area is considered urban. A comparison between all three methods revealed that in cases of closely spaced uniformly distributed points Averaging should be used because it gives the smallest mean value of the standard deviaion (3.0 dB). Furthermore, with the standard deviation of the error ranging from 2 to 11 dB interpolation techniques can be used for signal level prediction purposes themselves.",
keywords = "signal level interpolation, signal level prediction, propagation model, coverage area, mobile communications",
author = "Wenche Backman",
note = "Master's Thesis Helsinki University of Technology. Department of Electrical and Communications Engineering CA: TTE",
year = "2003",
language = "English",
publisher = "Helsinki University of Technology",
address = "Finland",
school = "Helsinki University of Technology",

}

Backman, W 2003, 'Error correction of predicted signal levels in mobile communications: Master's thesis', Master Degree, Helsinki University of Technology, Espoo.

Error correction of predicted signal levels in mobile communications : Master's thesis. / Backman, Wenche.

Espoo : Helsinki University of Technology, 2003. 66 p.

Research output: ThesisMaster's thesisTheses

TY - THES

T1 - Error correction of predicted signal levels in mobile communications

T2 - Master's thesis

AU - Backman, Wenche

N1 - Master's Thesis Helsinki University of Technology. Department of Electrical and Communications Engineering CA: TTE

PY - 2003

Y1 - 2003

N2 - The aim of this work is to combine measured and predicted signal levels in order to improve the accuracy of a propagation model. In addition to inserting measured signal levels in the predicted coverage area matrix, new values og signal level are interpolated around the measured ones and inserted in the matrix. Three interpolation techniques have been evaluated: Moving Least Squares (MLS), Delaunay Triangulation and Averaging. The first two methods interpolate signal level values over a wide area with samples from only a few measurement points, while averaging requires closely spaced, evenly scattered points. The purpose of this work has been to build up a database of signal levels for mobile terminal location. In addition, the improved coverage area predictions can be used for network planning purposes. The standard deviation is of primal interest when the accuracy of the interpolated signal levels is assessed because a small standard deviation proves the consensus between prediction and measurementin signal level changes. For the propagation model used in this work, the mean of the standard deviation of the error for seven test cases is 5.6 dB, compared to 6.1 dB with the MLS method and 5.1 dB with Triangulation. In cases where there were no shadowing obstacles present and there were smooth changes in the signal level, the MLS method gave better results while Triangulation performed better in areas with abrupt changes in the signal level. Hence, MLS should be used for error correction in rural and suburban areas while Triangulation performs better in urban areas, metropolises and probably also indoors. The test area is considered urban. A comparison between all three methods revealed that in cases of closely spaced uniformly distributed points Averaging should be used because it gives the smallest mean value of the standard deviaion (3.0 dB). Furthermore, with the standard deviation of the error ranging from 2 to 11 dB interpolation techniques can be used for signal level prediction purposes themselves.

AB - The aim of this work is to combine measured and predicted signal levels in order to improve the accuracy of a propagation model. In addition to inserting measured signal levels in the predicted coverage area matrix, new values og signal level are interpolated around the measured ones and inserted in the matrix. Three interpolation techniques have been evaluated: Moving Least Squares (MLS), Delaunay Triangulation and Averaging. The first two methods interpolate signal level values over a wide area with samples from only a few measurement points, while averaging requires closely spaced, evenly scattered points. The purpose of this work has been to build up a database of signal levels for mobile terminal location. In addition, the improved coverage area predictions can be used for network planning purposes. The standard deviation is of primal interest when the accuracy of the interpolated signal levels is assessed because a small standard deviation proves the consensus between prediction and measurementin signal level changes. For the propagation model used in this work, the mean of the standard deviation of the error for seven test cases is 5.6 dB, compared to 6.1 dB with the MLS method and 5.1 dB with Triangulation. In cases where there were no shadowing obstacles present and there were smooth changes in the signal level, the MLS method gave better results while Triangulation performed better in areas with abrupt changes in the signal level. Hence, MLS should be used for error correction in rural and suburban areas while Triangulation performs better in urban areas, metropolises and probably also indoors. The test area is considered urban. A comparison between all three methods revealed that in cases of closely spaced uniformly distributed points Averaging should be used because it gives the smallest mean value of the standard deviaion (3.0 dB). Furthermore, with the standard deviation of the error ranging from 2 to 11 dB interpolation techniques can be used for signal level prediction purposes themselves.

KW - signal level interpolation

KW - signal level prediction

KW - propagation model

KW - coverage area

KW - mobile communications

M3 - Master's thesis

PB - Helsinki University of Technology

CY - Espoo

ER -

Backman W. Error correction of predicted signal levels in mobile communications: Master's thesis. Espoo: Helsinki University of Technology, 2003. 66 p.