A Practical Strategy For Detecting Multiple Gross Errors

Abstract

Survey measurements are subject to random, systematic and gross errors. In practice, it is assumed that measurements are random variables, follow the normal distribution and have redundancy. The method of least squares estimation (ESE) is commonly used to process the redundant measurements. In the presence of gross errors, the results of ordinary ESE are corrupted. Consequently, an interactive post-LSE technique of robustified LSE (RLSE) is introduced for the detection of multiple gross errors in uncorrelated surveying data. In RESE. the locations and magnitudes of such errors are recovered simultaneously, and their effect on the solution are areatlv reduced.