实时测控系统数据分析与处理方法研究  

Data Analysis Processing Methods in Real-Time Measurement and Control System

作　　者：王鹏宇[1] 卞光浪[1] 秦刚 WANG Pengyu;BIAN Guanglang;QIN Gang(No.91550 Troops of PLA,Dalian 116023)

机构地区：[1]中国人民解放军91550部队

出　　处：《舰船电子工程》2020年第1期175-179,共5页Ship Electronic Engineering

摘　　要：飞行器测控系统在实时数据处理过程中,由于被测目标飞行时间长、传感器参数多、数据传递频率高,会产生大量处理结果数据。同时,由于受被测目标系统内部及传输信道噪声影响,数据结果质量与理想情况偏差较大。为了利用海量非理想实时数据快速的完成飞行器工作状态结果判定,论文提出了数据清洗、特征选取、结果预测、模型评价方法。最后以Python语言为平台,完成数据的分析与处理,并对飞行器工作状态进行预测。通过仿真结果分析,验证了处理方法的有效性,并在实际工程中取得了良好效果。Aircraft measurement and control system will generate a lot of data in processing,because of the target has a long time flight,many sensor parameters and high data rate transmission.There is a big deviation between actual data quality with desired value due to noise interference in internal system and transmission channel.To make it easy to quickly complete working state prediction,this paper presents the methods of data cleaning,feature extraction,condition prediction and model-evaluation.At last we complete data analysis and processing based on Python language.Some simulations were made to confirm the effectiveness of processing methods which have got good application.

关 键 词：实时数据 数据清洗 特征选取 ROC曲线 

分 类 号：V557[航空宇航科学与技术—人机与环境工程]