Scholarlink

To improve the accuracy and robustness of predicting aerial target trajectory under abnormal detection information, a trajectory prediction model of aerial target based on multi-head attention-LSTM is developed. The abnormal data is identified and deleted based on the moving boundary of the target, and the missing values are interpolated using a bidirectional prediction interpolation mechanism based on ARIMA. The multi-head attention module is designed to enhance the prediction performance by selectively focusing on the historical time series and the current state. The simulation results demonstrate that the multi-head attention-LSTM model can effectively predict the aerial target trajectory even in the presence of data anomalies and missing data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-head Attention-LSTM-Based Aerial Target Trajectory Prediction Under Abnormal Detection Information

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Xizhong Yang,
  • Tongle Zhou

摘要

To improve the accuracy and robustness of predicting aerial target trajectory under abnormal detection information, a trajectory prediction model of aerial target based on multi-head attention-LSTM is developed. The abnormal data is identified and deleted based on the moving boundary of the target, and the missing values are interpolated using a bidirectional prediction interpolation mechanism based on ARIMA. The multi-head attention module is designed to enhance the prediction performance by selectively focusing on the historical time series and the current state. The simulation results demonstrate that the multi-head attention-LSTM model can effectively predict the aerial target trajectory even in the presence of data anomalies and missing data.