@Monica Randall the concept is solid and yeah, for 2-3 second gaps it can definitely work, but the drift reality is a bit sobering. Consumer-grade MEMS accelerometers (which is what's on the sensors module) accumulate position error fast because you're double-integrating — acceleration → velocity → position — so any small bias or noise gets squared. At walking speed you're probably looking at 0.5-2 meters of drift over 3 seconds which is actually not terrible given the GNSS module's 1.5m base accuracy anyway. At vehicle speed it gets worse faster because the dynamic range is higher and small angle errors in your coordinate frame start mattering more.
The magnetometer on the sensors module is actually your friend here though. If you're fusing accel for dead reckoning, having a stable heading reference from the mag means you're not compounding both position and direction error simultaneously — that combo (accel + mag) is way more useful than accel alone for short gaps.
The 18Hz fix rate on the GNSS module is also working in your favor — you're getting a correction opportunity every 55ms when you do have a fix, so your Kalman filter (or whatever fusion approach you're using) can re-anchor pretty aggressively and bleed off the accumulated drift quickly after recovery.
