In this paper, we propose a novel neural network based state constrained integral sliding mode (NN-SCISM) control algorithm for nonlinear system with partially unknown dynamics in presence of time-varying constraints. In particular, the drift term characterizing the system dynamics is estimated by using a two-layer neural network, whose weights are adjusted according to adaptation laws designed relying on stability analysis. Thanks to a sliding variable which varies depending on the minimum distance between the system state and the current closest constraint, the control algorithm is able to drive the system state to a desired target state, while avoiding the forbidden states contained in the time-varying set delimited by the constraints. The proposal has been theoretical analysed and assessed in simulation.
Neural network based integral sliding mode control of systems with time-varying state constraints
Sacchi, Nikolas;Vacchini, Edoardo;Ferrara, Antonella
2023-01-01
Abstract
In this paper, we propose a novel neural network based state constrained integral sliding mode (NN-SCISM) control algorithm for nonlinear system with partially unknown dynamics in presence of time-varying constraints. In particular, the drift term characterizing the system dynamics is estimated by using a two-layer neural network, whose weights are adjusted according to adaptation laws designed relying on stability analysis. Thanks to a sliding variable which varies depending on the minimum distance between the system state and the current closest constraint, the control algorithm is able to drive the system state to a desired target state, while avoiding the forbidden states contained in the time-varying set delimited by the constraints. The proposal has been theoretical analysed and assessed in simulation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
