[teqc] MultiPath with L1 GPS receiver

James Johnson jjohnson at ucar.edu
Thu Feb 3 18:26:54 MST 2005


Dear Hannan,

I am taking about the wavelength of the multipath signal.  There are 
two factors that are important with multipath -- first is obviously the 
magnitude of the error in the C/A measurement.  These are the errors 
that corrupt the solution and cause position errors.  The second thing 
to consider is how rapidly the errors are changing.  Since multipath is 
an interference phenomena that depends on the excess path length of the 
reflected signal, the induced error will be cyclical with time.  If the 
frequency of this cyclical error is relatively fast, then one can 
compensate by using carrier phase smoothing and as long as the 
multipath can be averaged out before the accumulated ionosphere error 
becomes large, he phase smoothing will work well.  If the multipath 
signal has a long wavelength in time, then phase smoothing does not 
help.

-Jim



On Feb 3, 2005, at 6:17 PM, Hannan Sadar wrote:

> Thanks for the fast replay.
>
> I did not understand the last part of Jim's answer:
> "I think you will find quite a bit of multipath with long wavelengths
> (2-5 min), especially if you are using a cheap patch antenna."
>
> The wavelengths is a fix to ~0.19m. I would be glad if you could
> elaborate about this.
>
> Thanks
>
> Hannan
>
>
> On Thu, 3 Feb 2005 18:06:51 -0700, James Johnson <jjohnson at ucar.edu> 
> wrote:
>> Dear Hannan,
>>
>> Teqc requires dual frequency data in order to a multipath estimate.
>> This is because teqc looks at a long term difference between the phase
>> and code measurements while removing the ionosphere.  In principle, 
>> you
>> can do this with single frequency data over a short period of time, 
>> but
>> after some time the L1 phase and C/A measurements will diverge due to
>> the ionosphere (phase advance on L1 and group delay on C/A).  What you
>> want to look at is a running average of C/A - L1*lambda where C/A is
>> the pseudo range measurement, L1 is the phase measurement in cycles,
>> and lambda is the wavelength of L1.  If you use a running average of a
>> 10 min, you should be able to see the multipath signal, an you could
>> look at the RMS of this value (with steps removed caused by possible
>> cycle slips).  Unfortunately teqc does not do this.
>>
>> I think you will find quite a bit of multipath with long wavelengths
>> (2-5 min), especially if you are using a cheap patch antenna.
>>
>> Hope this helps,
>> -Jim
>>
>> On Feb 3, 2005, at 5:53 PM, Hannan Sadar wrote:
>>
>>> Hi,
>>>
>>> I have what is consider low cost GPS receiver and i would like to 
>>> find
>>> the multipath i was having on my measurements. I have psuedorange,
>>> phase and Doppler of L1. I would to know if teqc could find the
>>> multipath on of the measurements of teqc needs P1 data to do some
>>> averaging if I remember correctly.
>>>
>>> Thanks
>>>
>>> Hannan
>>> _______________________________________________
>>> teqc mailing list
>>> teqc at ls.unavco.org
>>> http://ls.unavco.org/mailman/listinfo/teqc
>>
>> _______________________________________________
>> teqc mailing list
>> teqc at ls.unavco.org
>> http://ls.unavco.org/mailman/listinfo/teqc
>>
> _______________________________________________
> teqc mailing list
> teqc at ls.unavco.org
> http://ls.unavco.org/mailman/listinfo/teqc



More information about the teqc mailing list