Focusing on the possibility for observed results with false or erroneous information of a sensor to affect data fusion estimation
an operator measuring consistency of the observed results of a multi-sensor is introduced. Before data fusion estimation
the consistency checking on the observed results is taken to identify and remove the observations of failed sensors
and obtain the consistent estimating. Based on the Hampel gross error identifier
the consistency measuring operator is able to take both self-support and mutual support between the sensors into account
to obtain the consistency information of observations. The simulation shows the simplicity and effectiveness for identifying false and erroneous observed results
eliminating the failed sensors
evaluating the quality of each sensor observations
and providing data fusion estimation with reliable consistent sensor group.