TY - JOUR
T1 - Three-dimensional path-following control for AUVs with unknown sideslip and attack angles
T2 - A predefined-time control approach
AU - Liu, Jian
AU - Wen, Guanghui
AU - Wu, Yongbao
AU - Li, Yiming
AU - Sun, Changyin
N1 - Publisher Copyright:
© 2025 Elsevier Ltd
PY - 2026/1
Y1 - 2026/1
N2 - This article investigates three-dimensional path-following control for autonomous underwater vehicles (AUVs) within a predefined-time (PT) framework, with a particular focus on addressing singular attitudes, unknown sideslip/attack angles, and the unwinding problem. A six-degree-of-freedom quaternion-based AUV model is first established to avoid Euler-angle singularities, followed by the derivation of roll-coupled path-following error dynamics. Then, a predefined-time extended state observer (PTESO) is proposed to estimate sideslip and attack angles without velocity measurements, enabling a PTESO-based line-of-sight guidance law. Furthermore, a new PT controller integrating a time-varying function and an anti-unwinding potential function is developed. This ensures unwinding-free performance throughout the control process, while also achieving PT convergence with smoother transient responses, reduced initial control magnitudes, and tighter bounds on settling time estimation. Simulation results confirm the effectiveness and advantages of the proposed method.
AB - This article investigates three-dimensional path-following control for autonomous underwater vehicles (AUVs) within a predefined-time (PT) framework, with a particular focus on addressing singular attitudes, unknown sideslip/attack angles, and the unwinding problem. A six-degree-of-freedom quaternion-based AUV model is first established to avoid Euler-angle singularities, followed by the derivation of roll-coupled path-following error dynamics. Then, a predefined-time extended state observer (PTESO) is proposed to estimate sideslip and attack angles without velocity measurements, enabling a PTESO-based line-of-sight guidance law. Furthermore, a new PT controller integrating a time-varying function and an anti-unwinding potential function is developed. This ensures unwinding-free performance throughout the control process, while also achieving PT convergence with smoother transient responses, reduced initial control magnitudes, and tighter bounds on settling time estimation. Simulation results confirm the effectiveness and advantages of the proposed method.
KW - Extended state observer
KW - Path-following control
KW - Predefined-time control
KW - Unit quaternion
UR - https://www.scopus.com/pages/publications/105019533015
U2 - 10.1016/j.automatica.2025.112651
DO - 10.1016/j.automatica.2025.112651
M3 - 文章
AN - SCOPUS:105019533015
SN - 0005-1098
VL - 183
JO - Automatica
JF - Automatica
M1 - 112651
ER -