I did some research as an undergrad in a mobile robotics lab at Universidade de São Paulo, and we had a super nice XSENS IMU for an autonomous car project, which did sensor fusion from GPS, gyroscope, accelerometer and magnetometer data (the last one was kept disabled for most of the time, as it’s common for ground vehicles because of interference). It provided incredible positioning accuracy even when losing GPS coverage, but it was very (!) expensive. This was in ~2010 and since then, many things have changed. XSENS still sells pretty expensive boxes, but MEMS technology evolved a lot and it’s much more affordable and popular today because of smartphones and drones.

I found an article one of these days comparing a few inexpensive IMUs and was really impressed with how good the results/specs seem to be. The article is from 2015, so things probably evolved a lot since then, but I sticked to it and got an MPU-9250. I want to integrate it with a custom motorcycle display I’m building, and see how far I can go with kalman filters. I’ll share the progress here.