The integration of terrain matching and inertial navigation is a crucial autonomous navigation method for low and medium altitude vehicles,but the traditional method has the following problems:the deformation of the elevation model due to the accumulation of inertial navigation errors restricts the terrain matching accuracy,the two-dimensional terrain matching has a large amount of computing power,and the positioning usability is poor after the terrain matching fails.A terrain matching navigation method based on frequency modulated continuous wave laser radar is proposed.Elevation model is constructed assisted by laser velocimetry,which reduces the deformation of the elevation model to improve the matching accuracy and reduces the correlation between the matching result and inertial navigation.A terrain matching algorithm integrating scale-invariant feature transform(SIFT)and stepwise verification is designed to enhance the real-time performance of the matching.The laser velocimetry is used to assist the correction of inertial navigation,which improve the positioning availability when the terrain matching fails.The proposed terrain matching navigation method is tested based on real terrain data of a typical scenario.The result shows that the terrain matching positioning accuracy of the elevation model based on laser velocity measurement is improved by 36.6%compared with that based on inertial navigation.The proposed terrain cardinality initialized random sample consensus terrain feature mismatch detection algorithm(TCI-RANSAC)reduces the verification time by 55.6%compared to the RANSAC algorithm.Furthermore,the introduction of the laser velocity information improves the integrated positioning accuracy by 65.9%.In conclusion,the proposed terrain matching navigation method has high engineering application value.
air vehicleFMCW LiDARterrain matching navigationintegrated navigationelevation model constructionfeature matching verification
维普
万方数据
