A novel digital twin model based bio-heuristic sliding mode control algorithm for trajectory tracking control of USV in the presence of complex marine environment disturbance

•USV’s trajectory tracking controller is designed based on digital twin model•Digital twin model of USV is established based on a novel AFF-RLS algorithm•Adaptive forgetting factor is proposed to flexibly adjust weights for historical data•Bio-heuristic function is presented to reduce abrupt changes...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Robotics and autonomous systems Ročník 195; s. 105219
Hlavní autoři:	Dong, Zaopeng, Lu, Sihang, Hu, Zhihao, Liu, Wangsheng, Ding, Yilun, Liu, Yuanchang
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	Elsevier B.V 01.01.2026
Témata:	Adaptive forgetting factor recursive least square (AFF-RLS) Digital twin model Radial basis function neural network (RBFNN) Sliding mode control (SMC) Trajectory tracking control Unmanned surface vessel (USV) Radial basis function neural network (RBFNN) Unmanned surface vessel (USV) Sliding mode control (SMC) Adaptive forgetting factor recursive least square (AFF-RLS) Trajectory tracking control Digital twin model
ISSN:	0921-8890
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	•USV’s trajectory tracking controller is designed based on digital twin model•Digital twin model of USV is established based on a novel AFF-RLS algorithm•Adaptive forgetting factor is proposed to flexibly adjust weights for historical data•Bio-heuristic function is presented to reduce abrupt changes in controller’s outputs•RBFNN algorithm is adopted to compensate for marine disturbances and modeling errors The trajectory tracking control problem of unmanned surface vessel (USV) under the complex marine environmental disturbance is discussed in this paper, and a novel digital twin model based bio-heuristic sliding mode control (SMC) algorithm integrated with a radial basis function neural network (RBFNN) disturbance compensation module is proposed. An adaptive forgetting factor, which varies with the prediction errors of state variables, is introduced and integrated into the recursive least squares (RLS) algorithm. Meanwhile, a digital twin model of USV is established by applying the proposed adaptive forgetting factor recursive least squares (AFF-RLS) algorithm, and utilizing state variable data and control commands. An improved bio-heuristic approximation function is presented to approach the virtual velocity control laws, avoiding abrupt change and jittering of the SMC algorithm designed based on the digital twin model. Set the angular and linear velocity variables as inputs, environment disturbances and modeling errors compensation as outputs, the RBFNN based integrated control compensator is presented, where the minimum parameter learning method is implemented by replacing the adjustment of neural network weights with parameter estimation to reduce redundant parameters. The effectiveness of the proposed algorithm is validated through extensive simulation experiments, demonstrating its robustness in real marine environment disturbance.
ISSN:	0921-8890
DOI:	10.1016/j.robot.2025.105219