A Driving Intention Inference Method Based on the Optimization of LSTM Network with Full Transfer Learning
Through the LSTM network optimization method based on complete transfer learning, a multivariate fractional-order gray model of driving intention was established, and driving intention was directly reasoned from the road conditions, which solved the problem of unreasonable intention reasoning and high-calculation burden in the existing technology, and achieved high-precision and efficient driving intention reasoning.
CN116739043BActive Publication Date: 2025-07-22CHANGCHUN UNIV OF TECH
View PDF 0 Cites 0 Cited by
Patent Information
- Application Number
- CN202310702577.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-14
- Publication Date
- 2025-07-22
- Estimated Expiration
- 2043-06-14
Smart Images
Figure CN116739043B_ABST
Abstract
The present invention belongs to the field of active safety technologies for electric vehicles, and specifically relates to a driving intention inference method based on the optimization of an LSTM network using complete transfer learning. The method includes the following steps: Step 1, establish a multi-variable fractional-order grey model for driving intention; Step 2, determine whether the source domain and the target domain are similar; Step 3, design and completely transfer the LSTM network; Step 4, perform optimization calculations on the fractional order to determine the driving intention. The present invention directly infers the driving intention from road conditions, with less interference information and higher accuracy; the introduction of the grey absolute correlation degree eliminates a large amount of probability calculations of data, greatly reducing the computational burden.
Need to check novelty before this filing date? Find Prior Art