1: \begin{abstract}

2: Terminal sliding mode (TSM) control algorithm and its non-singular

3: refinement have been elaborated for two decades and belong, since

4: then, to a broader class of the finite-time controllers, which are

5: known to be robust against the matched perturbations. While TSM

6: manifold allows for different forms of the sliding variable, which

7: are satisfying the $q/p$ power ratio of the measurable output

8: state, we demonstrate that $q/p=0.5$ is the optimal one for the

9: second-order Newton's motion dynamics with a bounded control

10: action. The paper analyzes the time-optimal sliding surface and,

11: based thereupon, claims the optimal TSM control for the

12: second-order motion systems. It is stressed that the optimal TSM

13: control is fully inline with the Fuller's problem of optimal

14: switching which minimizes the settling time, i.e. with

15: time-optimal control of an unperturbed double-integrator. Is is

16: also shown that for the given plant characteristics, i.e. the

17: overall inertia and control bound, there is no need for additional

18: control parameters. The single surface design parameter might (but

19: not necessarily need to) be used for driving system on the

20: boundary layer of the twisting mode, or for forcing it to the

21: robust terminal sliding mode. Additional insight is given into the

22: finite-time convergence of TSM and robustness against the bounded

23: perturbations. Numerical examples with different upper-bounded

24: perturbations are demonstrated.

25: \end{abstract}