... by accumulating height weights into float rather than uint8_t
these weights are all likely to be very low, so the "ends up at zero" case is likely to always be taken.
- min_flow_change is 0.04
- the value is in radians/second
I have an inkling this might originally have been "i" - i.e. accumulate across the three axes and then divide by the number of axes included in delta_height_m - but the code was subsequently changed to be include the abs_weight instead but the type was missed.
Peter Barker
b6d7107929