论文部分内容阅读
针对传统的线性预测方法精度较低的问题,提出基于非线性时间序列分析提出激光传感网络链路预测方法。对激光传感网络中的链路传输流量、信号强度和通信扩展范围等信息参量进行非线性时间序列分析,采用主成分分析方法实现对激光传感网络链路的特征信息融合处理,测量激光传感网络链路流量波动和信息流的主成分特征,利用灰度模型进行自适应学习,减少预测误差。最后进行仿真实验分析,结果表明,采用所提方法进行激光传感网络链路预测的准确度较高,收敛性较好,预测过程的开销较低,提高了链路吞吐量,性能优于传统方法。
Aiming at the problem of low accuracy of traditional linear prediction methods, this paper proposes a method of laser sensor network link prediction based on nonlinear time series analysis. Nonlinear time series analysis is carried out for the information parameters such as link transmission flow, signal intensity and communication extension range in the laser sensor network. The principal component analysis method is used to realize the characteristic information fusion of the laser sensor network link, Sense of network link flow fluctuations and information flow characteristics of the main components, the use of gray model adaptive learning to reduce the prediction error. Finally, the simulation experiment is carried out. The results show that the method proposed in this paper has higher accuracy, better convergence and lower cost of the forecasting process, improves the throughput of the link and has better performance than the traditional method.