Yes I’m still working on it and it did go somewhere. I posted the other day on the topic “Panda RVC.” I used the ESP32 as a test bed to gather data for testing the roll compensation algorithms (because it has bluetooth and I can use it with my phone and NTRIP). Was very interesting I learned that even when using the BNO in RVC mode, there is still latency and so the best IMU reading that corresponds with the GGA message is about 80-90 ms in the past. And I also found I needed little filtering on the roll value.
Anyway I’m currently implementing the corrected GGA/VTG stuff in Teensy now (for a purpose unrelated to AOG). When that is done in a week or two I can come back to the ESP32 panda code I based my work on and post it here. Obviously the ESP32 has no ether net so I stripped that out. It would be serial only. Is that something still useful to some? Also since the roll compensation is being outside of AOG there’s no way to control (without recompiling) whether pitch and roll are swapped, or if the direction of the values are reversed depending on the BNO’s orientation. As well AOG has no control over tractor reverse detection, so if the ESP gets it wrong, there’s no way for AOG to fix that without some additional code to both AOG and this modified PANDA sektch.