I’m trying to do range bias correction by making my own “RSL vs bias” table similar to application note APS011. However I’m not sure about what my antenna gain is in using Friis’ formula.
According to DWM1001 datasheet, table 11, it seems that if I position two modules parallel with each other, then the antenna gain is almost the same in every direction (dark blue curve). Even though this is the case, the gain still depends on angle of arrival, which I do not have access to when running real experiment.
Should I just go with a constant value like 0 or -1 or -2 for antenna gain in Friis’ formula just because it’s close enough? Or am I looking at the wrong place?
The correction is to correct for errors related to different RF power at the DW1000 input pins.
This means the power that you need to use as the input to your lookup table is the power after allowing for the antenna gain not the power before.
If the table is in terms of power after antenna gain and the power measurement you get is power after antenna gain then you don’t need to worry about antenna gain, the data in the table is independent of the antenna.
I’m sure that in the real world the antenna gain will have some impact on the data, RF is never that simple. But it’s going to be a secondary effect that will be small in comparison to other error sources in the system.
The only reason to need the antenna gain is if you wish to calculate the received power based on the transmit power and measured range.
This would make your process:
Measure range.
Calculate power based on range, transmit power and antenna gain.
Look up power in table and correct range.
But if you are assuming constant antennas and constant transmit powers then there is a fixed relationship between power and measured distance. So why not skip the middle step and create a look up table based on measured range rather than calculated power?
Yes you’re right. The process you described is how I imagined to do this. It’s just the table provided in APS011 is “RSL vs bias” so I wanted to keep it consistent.
It seems to me that using Friis’ formula (your process above) is the only way of doing this because there’s no off-the-shelf function that can return Rx power measured by DW1000 chip?
It is possible to calculate an estimate of the total received power, the user manual section 4.7.2 gives the method. But it’s not great, it tends to give a very noisy result.
But why do you need the power? If power is purely a function of distance and you’ve already measured distance then ignore power completely and create a table of measured range to range correction needed.