Robust neural network‐based tracking control for unmanned surface vessels under deferred asymmetric constraints

Abstract

In this article, the constraints on unmanned surface vessels are imposed after a certain time after the system operation. Based on the shift function, this article proposes an asymmetric barrier Lyapunov function and achieves the control scheme with deferred and asymmetric full-state constraints. In addition, radial basis function neural network and an antiwindup compensator are adopted for the problems of uncertainties and input saturation. The simulation results demonstrate the feasibility of the proposed control strategies.

Learning‐based robust control methodologies under information constraints

“…In addition, the command filtered technology is applied to the adaptive neural network design framework. Moreover, the authors in Reference 16 developed a control scheme with deferred and asymmetric full‐state constraints based on an asymmetric barrier Lyapunov function. In addition, radial basis function neural network and an antiwindup compensator are adopted for the problems of uncertainties and input saturation.…”

Adaptive integral terminal super‐twisting with finite‐time convergence for an unmanned surface vehicle under disturbances

“…Furthermore, consider the dynamic system, σ = a(𝜘, t) + b(𝜘, t)u, (5) where 𝜎 = 𝜎(𝜘, t) ∈ R is a sliding variable, which is designed such that the nonlinear system (1) reaches its desired dynamics in the SM 𝜎 = 0. In addition, the relative degree of (…”

“…Lemma 2 (24). Consider the dynamic system (5) where the functions a(𝜘, t), b(𝜘, t) satisfy Assumptions 1 and 2 for unknown parameters 𝛿 1 , 𝛿 2 , 𝛾 1 ∈ R + . Then, for any initial conditions 𝜘(0), 𝜎(0), there exist a finite-time 0 < T F and a parameter 𝜖 ∈ R + , so that a real 2-sliding mode, that is, |𝜎| ≤ 𝜂 1 and | σ| ≤ 𝜂 2 , is established ∀t ≥ T F , in presence of bounded additive (9) and multiplicative (6) perturbations with unknown boundaries, via the AST control…”

“…For fully‐actuated vessels, trajectory tracking has been widely studied. Different techniques have been applied, such as data‐driven control, 3,4 neural‐network‐based control, 5,6 and extended state observer‐based control 7 . Nevertheless, given the theoretical and practical challenge it represents, the trajectory tracking scenario is still a research interest for underactuated USVs 2,8 …”

Singularity avoidance adaptive output‐feedback fixed‐time consensus control for multiple autonomous underwater vehicles subject to nonlinearities

“…Thus, model-based control strategies are no longer feasible once there exist unknown parameters in the controlled system. Particularly, approximation-based approaches, such as neural networks (NNs) or fuzzy logic systems (FLSs)-based adaptive control, provide an effective solution to approximate the various kinds of uncertainties including unknown nonlinearities, [8][9][10][11][12][13] unknown external disturbances, 14,15 and unmodeled dynamics. 16 An NN target tracking control strategy for AUVs was proposed to compensate model uncertainties in Reference 17. Recently, a backstepping-based adaptive fuzzy tracking control approach was developed for AUVs with output constraints, which solved the issue of model uncertainties in Reference 18, to name just a few.…”