轨道交通GPS错误数据检测算法研究及验证

对某线两车站间轨道的58601组GPS测量数据的分析发现,数据存在测量过密或重复记录、测量过疏或忘记记录、往返测量、测量或记录错误等4种错误模式。为此提出对应的4种算法进行检测判断:针对前2种错误模式,根据相邻两点之间的距离来判断;针对错误模式3,根据当前点和其后第2点之间的间隔距离来判断;针对错误模式4,根据相邻线段的角度变化来判断。利用Matlab编程实现了算法,以58601组数据为例进行计算和确认,并与人工判断相比较。结果表明:该算法是有效的,检出率为100%,大幅压缩了检测时间,既减少了工作量,也保证了判断的正确性。

An Algorithm for Finding Errors in the Surveyed GPS Data of Rail Transit and Its Validation

Through the analysis of the 58 601 surveyed GPS data between two railway stations,four error patterns,including repetitive measurement,missing points,circular measurement and false measurement are found,and four algorithms are put forward to detect the corresponding error.The distance between two adjacent points is used to judge the two former error patterns.The distance between the current point and the next second point is used to judge the third error pattern.The change of the angles of the two adjacent lines is used to judge the fourth error pattern.The algorithms are realized using Matlab programming and the 58 601 surveyed data are used to validate the algorithms.Compared to the manual judgment,the algorithms are effective and the detection rate of the algorithms is 100%,which not only can decrease the workload markedly,but also insure the correction of the judgment.