期刊文献+

考虑波动因素的终端区空中交通流量预测仿真

Simulation of Air Traffic Flow Prediction in Terminal Area Considering Fluctuation Factor
下载PDF
导出
摘要 传统参数回归算法无法有效精确的对空中交通流量时序数据进行短期预测,为提升空域交通管理能力,提升流量短时预测的精确性与稳定性,将CV数据周期基础预测机制有机融入长短期记忆递归网络(LSTM)中,构建出CV-LSTM空中交通流量时序数据短时预测模型。模型首先采用WT算法对HBD数据进行优化分解,将时序数据分解为低频分量和噪声分量;然后采用离散数据周期分析的方式,通过分析数据的周期性与时空相似性,对高频噪声分量进行去噪处理;接着基于数据特征,构建出CV周期基础预测模型,完成数据的波动性预判与剔除;最后通过构建LSTM流量预测模型,在BP网络二次优化中,完成空中交通流量预测任务。多组流量预测模型仿真对比结果显示,在HBD数据集10min间隔的时间序列下,CV-LSTM模型预测曲线的追踪性能最好,且较其它五类基线模型相比,P、R和F_(1)指标整体分别平均提升了14.76%、15.55%和13.57%。综上,所构建的CV-LSTM空中交通流量时序数据短时预测模型具有最高的预测准确性以及较高的预测稳定性,在空中交通流量预测仿真中具有重要的研究价值。 Traditional parameter regression algorithm cannot effectively and accurately predict the time series data of air traffic flow in the short term.In order to improve the ability of airspace traffic management and the accuracy and stability of short-term flow prediction,this paper organically integrated the CV data period-based prediction mecha-nism into the Long-Short-Term Memory Recurrent Network(LSTM).The short-term prediction model of CV-LSTM air traffic flow time series data was constructed.Firstly,the HBD data were decomposed into low frequency components and noise components by using the WT algorithm,and then the discrete data cycle analysis was used to denoise the high frequency noise components by analyzing the periodicity and spatio-temporal similarity of the data;Then,based on the characteristics of the data,the basic prediction model of CV cycle was constructed to predict and eliminate the CVlatility of the data.Finally,by constructing the LSTM flow prediction model,the air traffic flow prediction task was completed in the secondary optimization of BP network.The simulation results show that the CV-LSTM mod-el has the best tracking performance in the 10 min interval time series of HBD data set,and compared with the other five baseline models,the P,R and F_(1) indicators are improved by 14.76%,15.55%and 13.57%,respectively.To sum up,the CV-LSTM air traffic flow time series short-term prediction model has the highest prediction accuracy and high prediction stability,and has important research value in air traffic flow prediction simulation.
作者 杨国洲 王云毅 YANG Guo-zhou;WANG Yun-yi(Air Traffic Control and Navigation College,Air Force Engineering University,Xi'an 710000 China)
出处 《计算机仿真》 2024年第11期6-10,177,共6页 Computer Simulation
关键词 流量预测 长短期记忆递归网络 周期波动 Traffic prediction LSTM network Periodic fluctuation
  • 引文网络
  • 相关文献

参考文献12

二级参考文献107

共引文献74

;
使用帮助 返回顶部