High-speed railway track multi-target fine tuning method driven by embedded physical neural network

The invention relates to the technical field of railway tracks, in particular to a high-speed railway track multi-target fine tuning method driven by an embedded physical neural network, which comprises the following steps of: 1) determining initial input data dimension and length according to track irregularity indexes; 2) establishing a potential feature expression layer, and performing multi-layer nonlinear low-rank transformation and feature activation on an input vector; 3) embedding a multi-track parameter constraint layer, and performing hard constraint according to a management value; 4) designing a weighted loss function of an embedded track characteristic and chord measurement formula; and 5) performing adaptive iterative optimization and scheme selection output. According to the method, multi-target fine adjustment of the high-speed railway track can be well carried out. 本发明涉及铁路轨道技术领域，具体地说，涉及一种嵌入物理神经网络驱动的高速铁路轨道多目标精调方法，其包括以下步骤：1)依据轨道不平顺指标确定初始输入数据维度和长度；2)建立潜在特征表达层，对输入向量进行多层非线性低秩变换和特征激活；3)嵌入多重轨道参数约束层，