Hi, I’m looking into the data outputted by the kraken, specifically the 360 degrees of DOA output.
The documentation says that each value is ’ DOA power output at X degrees ', what is the unit of this? RSSI power is in dB less that 0 and DOA output is always larger than 0, so it can’t be this. The sum of the DOA output isn’t consistent so it doesn’t look like its some sort of proportion at each angle of a whole value.
There seems to be a correlation between the confidence value and how large the max DOA output value is. I can’t seem to find the confidence value code (a previously posted link doesn’t work) so if I could be pointed to that it would be much appreciated.
Thank you,
Hi kraken forum,
Do we have any thoughts on this? The code is a bit opaque on what the doa output actually is and it would really help me if I could understand it,
Thank you
Sorry for the delay, was just confirming with our DFing algorithms expert. It’s a bit complicated and it varies per algorithm.
Bartlett - The output is power angular distribution
Capon - The output is SNR, or proportional to SNR
MEM - Like Bartlett but distorted. Probably not possible to have a unit.
MUSIC - It measures orthogonality between subspaces, so there is no physical quantity it can relate to.
So with MUSIC there are no units you can really use. But what you can be sure of is that a higher single skinnier peak is a better result compared to a flat graph or one with many peaks. The confidence value attempts to capture that into a single value.
But as mentioned before the confidence value is essentially depreciated, as we no longer use it in the mapping application, since the grid method works much so better to capture DOA result quality.
Thank you for your response, I hadn’t thought about the output being different between algorithms,
Jessie