New control method handles unknown nonlinear systems with input limits
A model-free controller achieves preset accuracy even under unknown saturation.
A team led by Dianrui Mu at Yanshan University has published a new control framework that tackles one of the hardest problems in nonlinear systems: maintaining performance when actuators hit unknown limits. Their paper, posted on arXiv, introduces a fully actuated manifold constraint-based output feedback controller that requires no system model and works for time-varying nonlinear dynamics with unknown input saturation.
The key innovation lies in generalizing existing constraint control methods—previously limited to linear manifolds—to arbitrary nonlinear manifolds. This allows the controller to guarantee a user-preset tracking accuracy within a finite or fixed time as long as the actuator is not saturated. If saturation occurs, the system gracefully degrades to a flexible accuracy, guided by an error-driven flexible constraint mechanism. The authors provide rigorous proofs and simulations on second-order and higher-order systems, demonstrating robustness and low computational complexity.
- Model-free design: no knowledge of system dynamics or input limits required
- Preset accuracy achieved in finite/fixed time under no saturation; flexible accuracy under unknown saturation
- Extends linear manifold constraint methods to general nonlinear manifolds with 22 pages of proofs and 12 figures
Why It Matters
Simplifies control of drones, robots, and autonomous vehicles dealing with unpredictable actuator limits.