To increase the accuracy in the local positioning system we are creating, we want to calibrate the antenna delay. Also, to calibrate the antenna delay, we want to create a reference device.
When creating a reference device, we use the methodology described in APS 014 paragraph 2.2.1. That is, we take three devices that we place in the corners of an equilateral triangle. The length of the sides of the triangle is known. Next, we compose a Euclidean matrix with these known actual distances.
Then we set the antenna delay in the devices to 0 and make a measurement between the devices using the TWR method, recording the measured values.
Then we have a question that we haven’t been able to answer yet. In APS 014, table No. 3 is presented as an example, which we are focusing on. In this table, it is not clear to us how the values RangeBias@162m, PRF=16MHz, Channel=were determined2 and RangeBias@7,914 m, PRF=16MHz, Channel=2, which affect the values of distances entered in the second Euclidean matrix.
Please explain how the values in the example in APS 014 were obtained, and how we should determine the RangeBias values when creating a reference device.
This number is an RSSI based correction factor, see section 3 of APS011.
The correct correction factor for your system will depend on the transmit power, the distance used and the antenna gain.
While you can use values in table 2 of APS011 my advice would be to characterise this correction factor for your system and use your own correction table.
This has the advantage that you can then define which range / RSSI you want to call a correction of 0. You can then make this zero point such that it matches the range you are using for your delay calibrations.
We know that this correction factor is RSSI based and depends on the transmit power, the distance used and the antenna gain. We had studied APS011 also. The question is that there are no RangeBias values in the table 2 APS011 that are given in table 3 in example in APS014. I would like to have explanations of how the values appeared in table 3 APS011 when RangeBias@162m, PRF=16MHz, Channel=2 and RangeBias@7.914m, PRF=16MHz, Channel=2. And we do not understand why in Table 3 APS014 two RangeBias values are subtracted from the measured distance value with an antenna delay equal to 0. It is not clear to us how the numbers +0.13 and -0.07 turned out in table 3 and why both of these values are taken into account when getting the values entered into the matrix.
Do you suggest, that we use our own values instead of the correction coefficients given in the table? Can you then tell us how to create our own correction table yourself?
How I created my own was I collected large amounts of data of UWB measured range vs truth between 8 different units and found the delay calibration values and RSSI correction factors that resulted in minimum average error for that data set.
I then assume the same RSSI correction factors for all units and so future units only need the antenna delay calibrated against a known unit.