Abstract: In UAV fixed-point flight verification, the real-time height measurement accuracy determines the angle verification error of PAPI (precision approach path indicator). Aiming at the large information error of fusion height caused by signal mutation of rotor UAV multi-airborne height sensor, a precise height measuring method based on FTKFT (fault-tolerant Kalman filter) is proposed consisting of a main Kalman filter and three sub-filters (GPS/barometric altimeter/radar altimeter). Each height sensor is filtered by Kalman filter， calculating the height estimation and the error as detecting signal and the second-order integration of inertial navigation in short time as reference signal， conducting state Chi-square test and residual Chi-square test with error detector. Finally， optimal data fusion of multi height sensors basing on output error of each height sensor is realized with weighting.Simulation and flight experiment show that this method is capable to achieve the expected measuring accuracy and real-time fault-tolerant performance.

Key words: rotor UAV, height measurement, Chi-square detection, false detection and isolation, data fusion