17-03-2014, 08:11 PM
Abstract—The accuracy of global positioning system (GPS)
based point determination in urban environment is mainly
corrupted by multipath and signal diffraction. The measured
carrier-to-noise-power-density ratio (C/N0) has proven to be
an excellent tool for the estimation of the random errors of the
observables. A weight model is proposed in order to scale the
positioning accuracy. The influence of parameters in the model
to positioning accuracy is carefully analyzed. Then these
parameters are determined. As compared to the positioning
solution without weight, the positioning accuracy increases by
9.85%, and 2.36% to SIGMA model.
Keywords—Navigation; Weighted model; BLUE
I. INTRODUCTION
In satellite navigation, multiple ranges from the user at
known locations to the satellites are obtained by measuring
the time it takes for a signal to reach the user receiver. And
the accuracy of GPS based point determination in urban
environment is mainly corrupted by multipath, signal
diffraction [1] and dilution of precision (DOP) [2]. In order
to improve satellite positioning accuracy, careful weighting
of the observation is needed.
Several approaches have been suggested for satellite
selection to improve the positioning accuracy over the years.
The most distinguished method is the RAIM algorithm,
which is based on self-consistency checks of GPS data. The
RAIM algorithm was originally developed for aviation
purposes [2] but it has also been found applicable for urban
pedestrian positioning [3]. Employing RAIM requires setting
of certain limits for different parameters to detect faulty
signals. User receivers usually select four suitable satellites
that have best geometry based on DOP [4]. The DOP
increases when the number of satellites decreases [5]. This is
unfavorable to the positioning accuracy [2]. So it only
applies to the circumstances that the ability of hardware is
poor previously. In [5], the weighted model of a triangular
fuzzy number is proposed and the three dimensional bias
averagely decrease about 3%. The SIGMA-?weight model
[6], [7] links the observable to the variance of the
observation. The signal distortion coincides with a deviation
of the actual from the template value [8]. And these
two models have been analyzed in [9]. The first weight
model is applied and tested indoor with high-sensitivity
receivers and the accuracy actually obtained was 40% better
than with equal weights [10].
C / N0
0 C / N
The , i.e. the ratio of the power level of the signal
carrier to the noise power in an 1 Hz bandwidth [11], is a
normalized measure of the signal-to-noise value. The 0 ,
elevation angle, URA and date of ephemeris have proven in
[12], to be a tool for the estimation of the random errors of
the observables. URA has a clear weighted factor in [13] and
date of ephemeris is analyzed in [11]. Due to the strong
correlation between the satellite elevation and the
[12], a template function can be determined to derive the
expected for an elevation angle and the specific
receiver-antenna combination [9].
0 C / N
C / N
0 C / N
0 C / N
The parameters can’t be achieved which calculate the
to an one-sigma error estimate with the formula in
[13]. So the new exponential weight model is presented
which analyze the influence to the positioning accuracy with
the information 0 in order to take random errors and
diffraction effects into account. Using this model, the
precision of the solution based on is different with
different parameters. The model presented is a further step
towards a more general weight model that uses measured and
estimated quality parameters of the observations.
based point determination in urban environment is mainly
corrupted by multipath and signal diffraction. The measured
carrier-to-noise-power-density ratio (C/N0) has proven to be
an excellent tool for the estimation of the random errors of the
observables. A weight model is proposed in order to scale the
positioning accuracy. The influence of parameters in the model
to positioning accuracy is carefully analyzed. Then these
parameters are determined. As compared to the positioning
solution without weight, the positioning accuracy increases by
9.85%, and 2.36% to SIGMA model.
Keywords—Navigation; Weighted model; BLUE
I. INTRODUCTION
In satellite navigation, multiple ranges from the user at
known locations to the satellites are obtained by measuring
the time it takes for a signal to reach the user receiver. And
the accuracy of GPS based point determination in urban
environment is mainly corrupted by multipath, signal
diffraction [1] and dilution of precision (DOP) [2]. In order
to improve satellite positioning accuracy, careful weighting
of the observation is needed.
Several approaches have been suggested for satellite
selection to improve the positioning accuracy over the years.
The most distinguished method is the RAIM algorithm,
which is based on self-consistency checks of GPS data. The
RAIM algorithm was originally developed for aviation
purposes [2] but it has also been found applicable for urban
pedestrian positioning [3]. Employing RAIM requires setting
of certain limits for different parameters to detect faulty
signals. User receivers usually select four suitable satellites
that have best geometry based on DOP [4]. The DOP
increases when the number of satellites decreases [5]. This is
unfavorable to the positioning accuracy [2]. So it only
applies to the circumstances that the ability of hardware is
poor previously. In [5], the weighted model of a triangular
fuzzy number is proposed and the three dimensional bias
averagely decrease about 3%. The SIGMA-?weight model
[6], [7] links the observable to the variance of the
observation. The signal distortion coincides with a deviation
of the actual from the template value [8]. And these
two models have been analyzed in [9]. The first weight
model is applied and tested indoor with high-sensitivity
receivers and the accuracy actually obtained was 40% better
than with equal weights [10].
C / N0
0 C / N
The , i.e. the ratio of the power level of the signal
carrier to the noise power in an 1 Hz bandwidth [11], is a
normalized measure of the signal-to-noise value. The 0 ,
elevation angle, URA and date of ephemeris have proven in
[12], to be a tool for the estimation of the random errors of
the observables. URA has a clear weighted factor in [13] and
date of ephemeris is analyzed in [11]. Due to the strong
correlation between the satellite elevation and the
[12], a template function can be determined to derive the
expected for an elevation angle and the specific
receiver-antenna combination [9].
0 C / N
C / N
0 C / N
0 C / N
The parameters can’t be achieved which calculate the
to an one-sigma error estimate with the formula in
[13]. So the new exponential weight model is presented
which analyze the influence to the positioning accuracy with
the information 0 in order to take random errors and
diffraction effects into account. Using this model, the
precision of the solution based on is different with
different parameters. The model presented is a further step
towards a more general weight model that uses measured and
estimated quality parameters of the observations.