Hello, I am developing an experimental indoor localisation application primarily using nRF5340-DKs. I have 3 of them in different corners of a room, in fixed positions, and one in the middle of the room, movable. The one in the middle scans for the RSSI values of the fixed ones and transmits them to a dongle connected to a laptop running a MATLAB GUI. The GUI needs to estimate the position of the movable nRF in the room and plot it on a map.
The RF transmission is done via ESB, on different pipes, with the PHY at 1M and the frequency at 2401MHz.
Besides using the Friis equation for trilateration, I am using RSSI fingerprinting, with a pre-trained decision tree map of measurements.
However, I am still not getting enough accuracy with both of these methods combined. What should I change in my application? Is there any way in which I could be implementing MCPD or ifft into ESB, as the nRF DM sample is using?
I did not want to center my application around nRF DM because I found it hard to implement in my network, on all 3 fixed devices and then into the dongle and GUI.
Thanks in advance!
