一种基于自注意力机制的AUV巡航速度估计方法
By extracting the temporal information of AUV data through a self-attention mechanism and an LSTM network, the problem of INS error accumulation when DVL is abnormal is solved, achieving high-precision AUV navigation estimation and improving navigation capability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HARBIN ENG UNIV
- Filing Date
- 2023-04-17
- Publication Date
- 2026-07-17
AI Technical Summary
In existing AUV navigation technologies, when the DVL output is abnormal for a long time, the INS error will accumulate, leading to a decrease in navigation accuracy. Furthermore, traditional methods fail to effectively utilize the temporal information of the data for correction.
A deep learning method based on self-attention mechanism is adopted. The temporal information of the data is extracted through LSTM network, and the correlation of data at different time points is calculated by using self-attention mechanism. Combined with INS data, high-precision navigation estimation is performed.
In cases where DVL is ineffective for extended periods, a high-precision navigation and positioning algorithm using the INS/DVL integrated navigation system was achieved, overcoming the spatial dependence of AUV navigation and improving navigation capabilities.
Smart Images
Figure CN117249822B_ABST