电子测量与仪器学报

王江,史元浩,郭正玉,田煜楷,韩天翔,李孟威.融合小波分解和 LSTM 的目标轨迹预测[J].电子测量与仪器学报,2023,37(1):204-211

融合小波分解和 LSTM 的目标轨迹预测

Target trajectory prediction by fusing wavelet decomposition and LSTM

DOI：10.13382/j.issn.1000-7105.2023.01.022

英文关键词:trajectory prediction recurrent neural networks wavelet decomposition long short term memory

基金项目:国家自然科学基金(72071183)、山西省自然科学基金(201901D111164)、山西省回国留学人员科研项目(2020-114)、中国高校产学研创新基金项目(2019ITA0cxy0023)资助

作者	单位
王江	1. 中北大学电气与控制工程学院
史元浩	1. 中北大学电气与控制工程学院
郭正玉	2. 中国空空导弹研究院
田煜楷	1. 中北大学电气与控制工程学院
韩天翔	1. 中北大学电气与控制工程学院
李孟威	1. 中北大学电气与控制工程学院

Author	Institution
Wang Jiang	1. School of Electrical and Control Engineering, North University of China
Shi Yuanhao	1. School of Electrical and Control Engineering, North University of China
Guo Zhengyu	2. China Airborne Missile Academy
Tian Yukai	1. School of Electrical and Control Engineering, North University of China
Han Tianxiang	1. School of Electrical and Control Engineering, North University of China
Li Mengwei	1. School of Electrical and Control Engineering, North University of China

摘要点击次数: 547

全文下载次数: 557

中文摘要:

随着目前空战武器装备的迅猛发展,对于高空高速大机动目标的轨迹预测越来越占据重要的战略地位。为了解决目前存在的目标轨迹预测不足的问题,本文提出了融合小波分解(wavelet decomposition, WD)和长短期记忆(long short term memory, LSTM)网络的模型来对机动目标的轨迹进行预测。首先,通过小波分解将输入的轨迹时间序列分解为 1 个低频分量(CD1)和 3 个高频分量(CA1,CA2,CA3)。然后,利用长短期记忆网络对时间序列处理的优势进行分量预测。最后,将分量预测结果进行重构并与原始轨迹进行对比验证,结果表明所提模型对于轨迹预测具有较高的精确度。为了排除实验结果的偶然性,本文用两组数据进行验证。通过对比实验显示,所提模型与其他两种模型相比预测误差更小。

英文摘要:

With the rapid development of current air combat weaponry, trajectory prediction for high-altitude, high-speed, large maneuver targets is occupying an increasingly important strategic position. In order to solve the current problem of insufficient target trajectory prediction, this paper proposes a model integrating wavelet decomposition (WD) and long short term memory (LSTM) network to predict the trajectory of maneuvering targets. First, the input trajectory time series is decomposed into one low frequency component (CD1) and three high frequency components ( CA1, CA2, CA3) by wavelet decomposition. Then, the component prediction is performed by taking advantage of the long short term memory network for time series processing. Finally, the component prediction results are reconstructed and compared with the original trajectories for verification, and the results show that the proposed model has high accuracy for trajectory prediction. In order to exclude the chance of experimental results, two sets of data are used for validation in this paper. The comparison experiments show that the proposed model has less prediction error compared with the other two models.

查看全文查看/发表评论下载PDF阅读器