[Prediction module] Why history_size needs saturating when judging obstacle's motion status?

[ra4-z]

In the function, Obstacle::SetMotionStatus(), WHY len and history_size used respectively?

When calculating avg_drift_x and avg_drift_y, both of them will be added for history_size times but divided by (len-1). WHY not (histoty-1) ? The same question on calculation of distance_std.

In my opinion, it would be good either only consideing len items in calculating avg_drift loop or just relying on history_size all the time.

Looking forward to your reply, thx :)