基于RBF-AR的船舶变形极短期预报

Short-term Ship Deformation Prediction Based on RBF-AR

摘要:

统一空间基准是海上作战平台实现精准探测打击的重要保证，而船体角变形的存在将严重影响空间基准的建立。针对这一问题，提出一种基于状态相依自回归（state-dependent auto-regressive, SD-AR）与径向基（radial basis function, RBF）神经网络的极短期变形预报方法，实现船体角形变的实时预报，为后续角变形的补偿提供依据。不同于传统的时间序列预报方法，该模型用一组RBF网络来逼近SD-AR模型中的函数系数，并采用一种结构化的非线性参数优化方法（structured nonlinear parameter optimization method, SNPOM）辨识该模型。基于该RBF-AR预报模型，给出了船舶变形预报算法设计并进行了仿真实验。实验结果表明，该方法在船体变形预测精度上优于传统时间序列预测方法，具有较好的应用前景。

Abstract:

The unified attitude reference is an important guarantee for the maritime combat platform to achieve accurate de-tection and strike. A prediction model combining state-dependent auto-regressive model with radial basis function neural networks is put forward for the problems that the existence of ship angular flexure makes it difficult to set up the unified attitude references. Unlike the current time series prediction methods, the model uses RBF neural net-works to approximate the parameters of SD-AR model, and the parameters of RBF neural networks are estimated with a structured nonlinear parameter optimization method, providing a basis for angular deformation compensation. According to the RBF-AR model, a design of theoretical algorithm and a mathematical simulation are carried out. The simulation results show that the prediction method is better than common time series prediction methods. The prediction model possesses best potential application in the field of ship angular flexure.