Rail transit vehicle health state evaluation and prediction method based on multi-modal data fusion

By quantifying the information value through the difference between information entropy and baseline information entropy, adaptively adjusting the temperature parameter of the attention mechanism, and dynamically adjusting the fusion weights of data from each modality, the problem of inaccurate prediction caused by changes in operating conditions in the health status assessment of rail transit vehicles is solved, and accurate and stable health status prediction is achieved.

CN121997276BActive Publication Date: 2026-06-09GUANGDONG HUANENG ELECTROMECHANICAL GRP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUANGDONG HUANENG ELECTROMECHANICAL GRP CO LTD
Filing Date
2026-04-07
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing methods for assessing the health status of rail transit vehicles cannot dynamically quantify and adjust the weights of each modality during data fusion under different operating conditions, resulting in inaccurate prediction results.

Method used

By collecting multimodal data and operating condition control signals, the difference between information entropy and baseline information entropy is calculated to quantify information value. The temperature parameter of the attention mechanism is adaptively adjusted using the coefficient of variation of information value, and the fusion weight of each modality data is dynamically adjusted. A cross-modal attention matrix is ​​constructed for weighted aggregation, and finally input into a long short-term memory network for prediction.

Benefits of technology

It achieves accurate and stable prediction of vehicle health status under different operating conditions, makes full use of complementary information from each mode, and reduces the interference of operating condition changes on the evaluation results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121997276B_ABST
    Figure CN121997276B_ABST
Patent Text Reader

Abstract

The application relates to the technical field of state prediction, in particular to a rail transit vehicle health state evaluation and prediction method based on multi-modal data fusion, which comprises the following steps: collecting multi-modal data and working condition control signals at any moment during the operation of a rail transit vehicle; constructing modal features and identifying the current working condition of the vehicle; calculating the information entropy of each modal feature, comparing the information entropy with the reference information entropy under the current working condition, and obtaining the information value of each modal data under the current working condition; calculating the variation coefficient of the information value of each modal data, and adjusting the temperature parameter of the attention mechanism according to the variation coefficient; weighting and aggregating each modal feature by using the temperature parameter and the information value to obtain multi-modal fusion features; and inputting the multi-modal fusion features at each moment into an evaluation and prediction model to output the vehicle health state. The technical scheme can realize accurate prediction of the vehicle health state.
Need to check novelty before this filing date? Find Prior Art