基于孤立森林算法的巡检机器人通信异常数据自动检测方法  

Automatic Detection Method of Abnormal Communication Data of Inspection Robot Based on Isolated Forest Algorithm

作　　者：刘杰[1] LIU Jie(Chongqing Creation Vocational College,Chongqing 402160,China)

机构地区：[1]重庆科创职业学院,重庆402160

出　　处：《信息与电脑》2023年第3期86-88,共3页Information & Computer

摘　　要：由于通信数据异常会影响巡检机器人对巡检任务数据的接收精度,因此利用孤立森林算法实现巡检机器人通信异常数据自动检测方法的优化设计。首先,在构建的通信信道环境下,采集巡检机器人通信数据,从通信流量、数据信噪比、数据包平均长度等方面,提取通信数据特征。其次,利用孤立森林算法,计算通信数据异常分值。最后,通过异常分值的分区比对和数据特征的匹配,以可视化的形式输出通信异常数据自动检测结果。通过性能测试实验发现,优化设计自动检测方法的误检率和漏检率均低于预设值,即优化设计方法具有较高的检测精度。Abnormal communication data will affect the receiving accuracy of the patrol robot to the patrol task data.Aiming at this problem,the isolated forest algorithm is used to realize the optimization design of the automatic detection method of the patrol robot's abnormal communication data.Firstly,in the constructed communication channel environment,collect the communication data of the inspection robot,and extract the characteristics of the communication data from the aspects of communication flow,data signal-to-noise ratio,average length of data packets,etc.Secondly,the isolated forest algorithm is used to calculate the abnormal score of communication data.Finally,the automatic detection results of communication abnormal data are output in visual form through the partition comparison of abnormal scores and the matching of data characteristics.Through the performance test experiment,it is found that the error detection rate and missing detection rate of the optimized design automatic detection method are lower than the preset value,that is,the optimized design method has high detection accuracy.

关 键 词：孤立森林算法 巡检机器人 通信异常 异常数据检测 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]