摘要
为解决水下无人潜航器长时间作业下航位推算误差累积难题,应用Informer时序预测模型,对UUV航行信息及推算航位内在关系进行深度挖掘和特征提取,预测目标未来时序下的航位误差趋势。进行对比实验和预测航迹可视化,实验表明Informer模型在预测复杂环境下航位推算误差问题上取得了较好效果,适用于复杂环境下UUV航位推算误差校正。
To solve the problem of accumulated navigation errors in underwater unmanned underwater vehicles during long-term operations,applying the Informer time series prediction model,deep mining and feature extraction of UUV navigation information and the inherent relationship between calculated navigation positions,predict the trend of heading error in the future time series of the target.Conduct comparative experiments and predict trajectory visualization,the experiment shows that the Informer model has achieved good results in predicting the problem of dead reckoning errors in complex environments,suitable for correcting UUV dead reckoning errors in complex environments.
作者
刘立伟
程卓
范学满
LIU Liwei;CHENG Zhuo;FAN Xueman(Jiangsu Institute of Automation,Lianyungang Jiangsu 222006;Naval Submarine Academy,Qingdao Shandong 266000)
出处
《软件》
2024年第8期135-138,共4页
Software
