| 傅 莉,邱鹏,姜冠武,范广兴.新型飞机雷达散射特性ARIMA-NARX预测[J].航空发动机,2026,52(1):103-111 |
| 新型飞机雷达散射特性ARIMA-NARX预测 |
| Prediction Analysis of Radar Scattering Characteristics of New Aircraft |
| |
| DOI：10.12482/ISSN.1672-3147.20230420002 |
| 中文关键词: 雷达散射截面 序列预测 优化定阶 差分自回归滑动平均模型 非线性自回归模型 耦合预测 |
| 英文关键词:RCS sequence prediction optimize the order auto-regressive integrated moving average model Nonlinear autoregres⁃
sive model coupling prediction |
| 基金项目:国家自然科学基金（61602321）资助 |
|
| 摘要点击次数: 1730 |
| 全文下载次数: 668 |
| 中文摘要: |
| 针对经典序列预测算法难以满足复杂雷达散射截面（RCS）序列的预测问题，提出了基于贝叶斯信息量准则的差分自回
归滑动平均模型-非线性自回归模型（ARIMA-NARX）耦合预测算法。采用商业软件求解所设计的新型飞机在不同频段下的
RCS，在ARIMA基础上引入贝叶斯信息量准则对模型进行优化定阶，采用优化后的ARIMA模型对RCS序列进行预测并计算预测
残差，采用最优NARX对残差值进行预测，将残差预测结果与ARIMA的RCS序列预测结果相结合，得到最终的RCS序列预测结果
从而构建ARIMA-NARX预测模型。以设计的新型飞机作为研究对象进行了不同极化条件下覆盖C、X和Ku频段的复杂RCS序列
预测设计与结果分析，结果表明：与传统预测算法相比，采用ARIMA-NARX预测算法得到的均方根误差、平均绝对误差和平均绝
对误差均减小60%以上。算法可推广应用于其他复杂RCS序列的预测。 |
| 英文摘要: |
| Aiming at the problem that the classical sequence prediction algorithm was difficult to meet the prediction of complex
Radar Cross-Section (RCS) sequences, an differential autoregressive moving average model-nonlinear autoregressive (ARIMA-NARX)
model coupling prediction algorithm based on Bayesian information criterion was proposed. Business software was used to solve the RCS of
the new designed aircraft in different frequency bands. Based on the ARIMA,the Bayesian Information Criterion(BIC)was introduced to
optimize the order of the model. The optimized ARIMA model was used to predict the RCS sequence and calculate the predicted residual.
NARX was used to predict the residual value,and the residual prediction results were combined with the RCS series prediction results of
ARIMA to obtain the final RCS series prediction results and construct the ARIMA-NARX prediction model. The prediction design and
result analysis of complex RCS sequence covering C, X and Ku bands under different polarization conditions were carried out with the new
designed aircraft as the research object. The results show that the root mean square error, mean absolute error and mean absolute error
percentage of the proposed ARIMA-NARX prediction algorithm are reduced by more than 60% compared with the traditional prediction
algorithm. The algorithm can be applied to the prediction of other complex RCS sequences. |
| 查看全文 HTML 查看/发表评论 下载PDF阅读器
|