受传感器精度限制及噪声干扰等不利因素的影响，当前无人艇在路径跟踪方面的精度与实时性尚存在明显不足。这些问题可能导致无人艇偏离预定路径，影响其执行任务的效率和安全性。针对这些问题，提出一种基于贝叶斯平滑-交替方向乘子法的无人艇路径跟踪方法。将无人艇路径跟踪问题视为一类非线性约束的非凸优化问题，融合交替方向乘子法（Alternating Direction Method of Multiplier，ADMM）和贝叶斯平滑器（Bayesian Smoothing，BS）理论，提出了贝叶斯平滑-交替方向乘子法（BS-ADMM），旨在有效解决迭代过程中的子问题。实验结果显示，相比于传统优化方法，该贝叶斯平滑-交替方向乘子法的路径跟踪计算时间提高约65%，误差降低15%。不仅提升了无人艇的路径跟踪精确度和稳定性，也为其在复杂海洋环境中的自主航行提供了有力支持。
Influenced by unfavorable factors such as sensor accuracy limitation and noise interference, the current unmanned surface vehicle’s accuracy and real-time performance in path-following still have obvious deficiencies. These problems may not only lead the USV to deviate from the predetermined path, but also affect the efficiency and safety of their missions. To address these problems, an unmanned surface vehicle path-following method based on Bayesian smoothing-alternating direction multiplier method is proposed. Considering the unmanned surface vehicle path-following problem as a class of nonconvex optimization problems with nonlinear constraints, the theory of alternating direction method of multiplier (ADMM) and Bayesian smoothing (BS) are fused, and in the iterative framework of ADMM, a new augmented Bayesian smoothing, aiming to solve the subproblems in the iterative process efficiently. Experimental results show that the path-following computation time of this Bayesian smoothing-alternating direction multiplier method is improved by about 65% and the error is reduced by 15% compared to the traditional optimization method. It not only improves the accuracy and stability of the unmanned surface vehicle's path-following, but also provides strong support for its autonomous navigation in complex marine environments.
2025,47(21): 65-72 收稿日期：2024-12-12
DOI：10.3404/j.issn.1672-7649.2025.21.012
分类号：U674
基金项目：国家自然科学基金青年科学基金资助项目（52301402）；上海市浦江人才资助项目（22PJ1405400）；上海交通大学深蓝计划资助项目（SL2022MS002）；广东省基础与应用基础研究基金资助项目（2022A1515110574）
作者简介：高源(2000-),男,硕士研究生,研究方向为机器学习算法及其在水面无人系统中应用
参考文献：
[1] 张韩西子, 倪海参, 石正坤, 等. 水面无人艇发展趋势及关键技术展望[J]. 舰船科学技术, 2024, 46(8): 108-111.
ZHANG H X Z, NI H S, SHI Z K, et al. Development trends and key technology prospects of unmanned surface vessels[J]. Ship Science and Technology, 2024, 46(8): 108-111.
[2] 李家良. 水面无人艇发展与应用[J]. 火力与指挥控制, 2012, 37(6): 203-207.
LI J L. Development and application of unmanned surface vehicles[J]. Fire Control & Command Control, 2012, 37(6): 203-207.
[3] 黎为. 基于预测控制的水面无人艇航迹跟踪方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2016.
[4] KATAYAMA H, AOKI H. Straight-line trajectory tracking control for sampled-data underactuated ships[J]. IEEE Transactions on control systems technology, 2013, 22(4): 1638-1645.
[5] CAO G, YANG J, QIAO L, et al. Adaptive output feedback super twisting algorithm for trajectory tracking control of USVs with saturated constraints[J]. Ocean Engineering, 2022, 259: 111507.
[6] ZHAO Y, QI X, INCECIK A, et al. Broken lines path following algorithm for a water-jet propulsion USV with disturbance uncertainties[J]. Ocean Engineering, 2020, 201: 107118.
[7] WU C, YU W, LIAO W, et al. Deep reinforcement learning with intrinsic curiosity module based trajectory tracking control for USV[J]. Ocean Engineering, 2024, 308: 118342.
[8] ZHONG W, LI H, MENG Y, et al. USV path following controller based on DDPG with composite state-space and dynamic reward function[J]. Ocean Engineering, 2022, 266: 112449.
[9] MOUSAZADEH H, JAFARBIGLU H, ABDOLMALEKI H, et al. Developing a navigation, guidance and obstacle avoidance algorithm for an Unmanned Surface Vehicle (USV) by algorithms fusion[J]. Ocean Engineering, 2018, 159: 56-65.
[10] 何静. 利用卡尔曼滤波预测船舶航行轨迹异常行为[J]. 舰船科学技术, 2017, 39(2): 16-18.
HE J. Predicting abnormal behavior in ship navigation trajectories using Kalman filter[J]. Ship Science and Technology, 2017, 39(2): 16-18.
[11] 董早鹏, 万磊, 廖煜雷, 等. 基于非对称模型的欠驱动无人艇路径跟踪控制[J]. 中国造船, 2016, 57(1): 116-126.
DONG Z P, WAN L, LIAO Y L, et al. Path tracking control of underactuated USV based on an asymmetric model[J]. Shipbuilding of China, 2016, 57(1): 116-126.
[12] FOSSEN T I, BREIVIK M, SKJETNE R. Line-of-sight path following of underactuated marine craft[J]. IFAC Proceedings volumes, 2003, 36(21): 211-216.
[13] 孙波, 熊常鹏. 常微分方程欧拉折线法的 Picard 迭代改进[J]. Advances in Applied Mathematics, 2023, 12: 3763.
SUN B, XIONG C P. Improvement of the Euler polygon method for ordinary differential equations using Picard iteration[J]. Advances in Applied Mathematics, 2023, 12: 3763.
[14] 王慧芳. 线性化乘子交替方向法求解稀疏组最小一乘模型[D]. 北京: 北京交通大学, 2017.
[15] GAO R, TRONARP F, SÄRKKÄ S. Iterated extended kalman smoother-based variable splitting for L_1 $-regularized state estimation[J]. IEEE Transactions on Signal Processing, 2019, 67(19): 5078-5092.
[16] BOYD S, PARIKH N, CHU E, et al. Distributed optimization and statistical learning via the alternating direction method of multipliers[J]. Foundations and Trends in Machine learning, 2011, 3(1): 1-122.
[17] HE B S, YANG H, WANG S L. Alternating direction method with self-adaptive penalty parameters for monotone variational inequalities[J]. Journal of Optimization Theory and Applications, 2000, 106: 337-356
[18] FOSSEN T I. Handbook of marine craft hydrodynamics and motion control[M]. Hoboken:John Willy & Sons Ltd, 2011.
[19] BELL B M, BURKE J V, PILLONETTO G. An inequality constrained nonlinear Kalman–Bucy smoother by interior point likelihood maximization[J]. Automatica, 2009, 45(1): 25-33.
