|
The existing robust estimation theory is based on the error classification philosophy, thinks the gross error is from the wrong measurement operatation, the processing mode always is adopted to weaken its influence (reducing weight) and even to eliminate it. In this thesis, by analyzing the physical mechanism of forming gross error, the author will prove many gross errors are actually normal measurement phenomenon, the information expressed by this kind of gross error is actually very valuable, the processing mode of reducing weight is not appropriate in existing researches, give the reasonable idea of processing this normal gross error. |
|
Keywords:gross error; robust estimation; least square method; regular error |
|