Mass-centric controller steers unmanned surface vehicles with AI
New Lyapunov-based neural controller shifts mass to navigate precisely.
Control of surface vehicles with dynamically changing mass distributions remains an open problem. Researchers from the University of Turku and the Federal University of Rio Grande do Norte propose a novel intelligent controller that uses mass distribution itself as a control input. The system shifts a sliding mass to alter the vessel's center of gravity, working alongside thrust for forward motion.
A Lyapunov-based nonlinear control scheme guides the vessel along a smooth path using a line-of-sight guidance law. Artificial neural networks estimate unmodeled dynamics and external disturbances, compensating for uncertainties. Simulation results demonstrate effective path-following with minimal error, as presented at the European Control Conference (ECC) 2026.
- Control inputs include thrust force and position of a sliding mass to shift overall mass distribution.
- Lyapunov-based nonlinear control scheme combined with line-of-sight guidance ensures path-following.
- Neural networks estimate unmodeled dynamics and disturbances, improving robustness in simulations.
Why It Matters
Enables more maneuverable unmanned vessels for missions like search-and-rescue or ocean monitoring.