The simulation currently assumes that you always start from a constant state, i.e. one where you are moving with a specific body / module command and have reached steady state. In the real world it is possible that new commands are given while existing commands aren't complete yet. This means that there are existing velocity / acceleration / jerk values. Ideally the transition profiles take those into account and work with them, instead of ignoring them.
The simulation currently assumes that you always start from a constant state, i.e. one where you are moving with a specific body / module command and have reached steady state. In the real world it is possible that new commands are given while existing commands aren't complete yet. This means that there are existing velocity / acceleration / jerk values. Ideally the transition profiles take those into account and work with them, instead of ignoring them.