We have a custom board with DW1000 and intend to implement two way ranging for distance estimation. We understand that a bias correction for the dependency of range error on received signal strength has to be done to improve the accuracy.
Can we reuse the table no : 2 (Relationship between RSL and range bias correction factor) of APS011 for this? If not, what would be the procedure to create such a table for our device? Also is there any particular reason why the TREK1000 source uses distance as input for the bias correction instead of power level?
Thanks in advance!
You could use the existing table but you will probably be able to obtain better results if you create your own for your equipment.
The procedure would be to take measure a series of points. At each point measure the signal level, the calculated range and the actual range. From this you can plot the range errors Vs the signal power.
The DW1000’s ability to measure signal level isn’t great. It works but the results are a little noisy. Distance measurements are a lot lower noise than signal power measurements.
Assuming direct line of sight and antenna heights/separations are such that you aren’t in a ground bounce fade then there will be a very strong correlation between signal level and distance.
All of which means you’ll get a more consistent and reliable result if you base your correction on range rather than signal power. Assuming LoS etc…
This could result in an incorrect correction when not LoS whereas a power level correction may be more accurate in that situation but these corrections are in the order of 5-10 cm, if it’s NLoS then the chances are your errors are bigger than that to start with.
To start off with , we shall be using the table provided by decawave using distance input for the bias correction. But going through the TREK source code, we came across a calculated_distance X 4, before finding the index of table corresponding to it. What could be the logic behind that?