Nonholonomic underactuated systems are typically modelled as highly nonlinear ones, which becomes obvious as the dimension of the system increases. To the author’s knowledge, so far there is no canonical form for the general nonholonomic systems, for example, free flying space robots. Therefore, even a proof of the controllability is made based on numerical computations, for example, see [49]. Furthermore, a motion planning requires high computational costs in dealing with complicated equations, for example see [82, 52].