[PDF] Gate Recurrent Unit Network based on Hilbert-Schmidt Independence Criterion for State-of-Health Estimation-论文阅读讨论-ReadPaper

摘要

State-of-health (SOH) estimation is a key step in ensuring the safe andreliable operation of batteries. Due to issues such as varying datadistribution and sequence length in different cycles, most existing methodsrequire health feature extraction technique, which can be time-consuming andlabor-intensive. GRU can well solve this problem due to the simple structureand superior performance, receiving widespread attentions. However, redundantinformation still exists within the network and impacts the accuracy of SOHestimation. To address this issue, a new GRU network based on Hilbert-SchmidtIndependence Criterion (GRU-HSIC) is proposed. First, a zero masking network isused to transform all battery data measured with varying lengths every cycleinto sequences of the same length, while still retaining information about theoriginal data size in each cycle. Second, the Hilbert-Schmidt IndependenceCriterion (HSIC) bottleneck, which evolved from Information Bottleneck (IB)theory, is extended to GRU to compress the information from hidden layers. Toevaluate the proposed method, we conducted experiments on datasets from theCenter for Advanced Life Cycle Engineering (CALCE) of the University ofMaryland and NASA Ames Prognostics Center of Excellence. Experimental resultsdemonstrate that our model achieves higher accuracy than other recurrentmodels.

展开全部

图表提取

论文十问

参考文献

被引用

提问与回答

暂无人提供速读十问回答