@Charles Hall yeah this is a genuinely interesting edge case and glad you're thinking about it before you're chasing phantom disconnects in the field.
The short answer is that Doppler at 120 km/h is basically a non-issue at 2.4GHz — you're looking at a frequency shift in the tens of Hz range, which is orders of magnitude below what the receiver's acquisition bandwidth cares about. The PHY won't even notice.
What actually matters at that kind of speed is the channel hopping behavior and how aggressively the connection parameters are tuned. BLE's adaptive frequency hopping is doing 37 data channels and it's remapping around interference in near-real-time, but if your connection interval is loose and the vehicle is moving through a multipath-heavy environment (think underpasses, buildings, other RF sources), you can get bursts of dropped packets that look like link instability but are really just the hopping algorithm catching up. Tightening the connection interval helps, and the STM32WB07 gives you enough control over that to tune it for a moving-node scenario specifically.
The other thing worth watching is antenna orientation — at vehicle-mounted speeds, especially if the node is rotating or the vehicle is changing heading relative to your ground station, you can get pattern nulls that hit faster than the AGC can compensate. If you're seeing intermittent drops rather than gradual RSSI degradation, that's usually the culprit. The firmware repo has some hooks you can use to log RSSI and packet error rate in real time, which I'd set up before your first mobile run so you have actual data to work with rather than guessing.
