Sliding-Mode Control
Sliding-mode control forces the state onto a stable surface and holds it there, giving strong robustness at the cost of chattering.
Robust Control by Switching
Sliding-mode control is a nonlinear method built for robustness. It defines a sliding surface in state space on which the system has desired, stable dynamics, then applies a switching control law that drives the state to that surface and keeps it there. Once on the surface, the system slides along it toward equilibrium, largely insensitive to disturbances and model error.
The two phases
- Reaching phase: the switching control forces the state toward the sliding surface from wherever it starts.
- Sliding phase: once on the surface, the state is constrained to it and follows the surface's designed dynamics.
- On the surface, behavior depends only on the surface definition, not on the uncertain plant details, which is the source of robustness.
Designing the surface
The sliding surface is chosen so that when the state is confined to it, the reduced-order dynamics are stable with the desired response, typically a linear combination of tracking error and its derivatives set to zero. Because the surface dictates behavior in the sliding phase, its design directly sets closed-loop performance.
Robustness and the matching condition
Sliding-mode control rejects any disturbance or uncertainty that enters through the same channel as the control input, the so-called matched uncertainty, provided the switching gain exceeds the disturbance bound. This strong invariance property is what makes the method attractive for systems with significant uncertainty.
Chattering and its cures
The main drawback is chattering: because ideal switching is infinitely fast but real actuators are not, the control oscillates rapidly around the surface, exciting high-frequency dynamics and stressing hardware. Practical fixes replace the hard switch with a smooth approximation inside a thin boundary layer, or use higher-order sliding modes such as the super-twisting algorithm, trading a little robustness for smoother action.
Sliding-mode control is valued wherever robustness to large, poorly modeled disturbances matters more than smoothness, from motor drives to robotics, and its stability is proven with a Lyapunov function built around the sliding surface.