U-blox TIM OEM Receiver (SiRF II) Tested

This page is under construction.

The Swiss firm u-blox AG donated a TIM OEM receiver for evaluation. The receiver is based on the SiRF-II chipset.
In this page the results of some tests are reported:

  1. In a so-called zero baseline setup the multipath of the raw C/A code and carrier were determined (see my multipath page for an explanation of the method). Also, various position solutions (such as the internal calculated position, stand-alone calculation and single difference calculation using raw and smoothed pseudoranges, were compared with the true position.

  2. Two short baseline tests, one in a multipath benign environment and one in a multipath heavy environment,  allowed me to gain more insight into the receiver behaviour under favourable conditions and under realistic conditions.

  3. Finally, the receiver tracking performance results in the urban canyon and in the woods are reported.

Experiment 1.

In the first experiment, I used Tri-M's Big Brother antenna on a small groundplane (25 cm dia) on top of my roof, with an unobstructed view over the horizon. The antenna was connected to a tri-way splitter with a DC coupling to my reference receiver, and the antenna input of the TIM, AC coupled via a 100pF capacitor. (the two receivers in a 'zero-baseline' setup). The TIM was set-up to generate raw measurement data in SiRF's binary format. The data was downloaded to the serial ports of my PC. About three hours of data were recorded on 29 May 2002, and processed.

1.1 C/A Code Multipath

The graph above shows the C/A code multipath magnitude of the TIM, connected to the Big Brother. The relation between the multipath (1 sigma value) and the satellite elevation is not very strong. The maximum value less than 7m, the average value for elevations between 20 and 90 deg about 2.5 m.

1.2 Carrier multipath

Carrier multipath is practically independent of elevation, with an average value of about 0.2 cycle (4 cm). This value is low enough to allow smoothing of the pseudoranges with the carrier ranges. The position calculation results below demonstrate the effectiveness of smooothing.

Similar to the earlier SiRF I based PS-1 receiver, double differenced carrier phase measurements of this zero baseline test are NOT constant in time and do NOT average to integer values. High accuracy position calculation based on double differenced carrier phase measurements is therefor not possible with this receiver (and probably not possible with any SiRF I or II based receiver).

1.3 Standalone position, raw data

The above scatterplots were produced by comparing calculated latitude, longitude and height with the true antenna position (known with an accuracy of a few cms'). More that two hours of data (about 8000 epochs) were processed. The raw pseudorange measurements as reported in binary message 28 were used in the standalone position calculation. Statistics for northing, easting and delta height were calculated:

statistic dLat [m] dLon[m] dH [m]
bias -0.7 0.1 3.8
sigma 2.9 1.9 5.5
95% 6.5 3.9 14.8

Horizontal error statistics are as usual, the bias in height is rather large however. This may be due to the 'Single Frequency Ionospheric Correction Parameters Anomaly' in the periopd from 28 May to 2 June 2002, as reported by the Department of The Air Force.

1.4 Standalone position, smoothed data

The above graphs show the effect of smoothing the raw pseudoranges with the carrier data, also present in message 28. The sudden jumps away from the true position, and subsequent convergence back to the true position result from the smoothing filter being reset either by a new satellite or by a carrier cycle slip.

The statistics are:

statistic dLat [m] dLon[m] dH [m]
bias -0.6 0.2 4.0
sigma 1.6 0.9 3.5
95% 3.8 2.0 11.0

The biases have not changed, but the standard deviations alomost halved in value. For this dataset smoothing proves to be effective, especially for the horizontal coordinates.

1.5 Single difference position, raw data

Although not visible in the above scattter plots, singls differencing the raw pseudoranges removes the biases almost completely, and lowers the standard deviations significantly. Please compare the statistics below with the statistics of the standalone calculations.

statistic dLat [m] dLon[m] dH [m]
bias -0.2 0.4 -0.1
sigma 2.1 1.6 3.8
95% 4.4 3.6 7.3

The height component improved by a factor of two, compared to the raw data standalone position statistics.

1.6 Single difference position, smoothed data

Single difference positioning with smoothed data improves the statistics with a factor of two:

statistic dLat [m] dLon[m] dH [m]
bias -0.3 0.4 0.0
sigma 0.7 0.6 1.6
95% 1.7 1.6 3.2

1.7 Statistics summary

All statistics are summarized below. And let the results speak for themselves!

Mode dLat [m] dLon [m] dHeight [m]
bias sigma 95% bias sigma 95% bias sigma 95%
ZD raw -0.7 2.9 6.5 0.1 1.9 3.9 3.8 5.5 14.8
ZD smooth -0.6 1.6 3.8 0.2 0.9 2.0 4.0 3.5 11.0
SD raw -0.2 2.1 4.4 -0.4 1.6 3.6 -0.1 3.8 7.3
SD smooth -0.3 0.7 1.7 -0.4 0.6 1.6 0.0 1.6 3.2

back to index