摘要：AbstractIn this article a fault detection algorithm for aircraft position measurements i proposed using redundant sensor information during landing scenarios. This work was developed in the framework of the VISION EU H2020 research project. The aircraft’s position can be determined via instrumental landing system, GPS and camera measurements. Considering these three sources a two out of three voting logic can be developed. After transforming the measured data sets to a common format two different methods are constructed to execute voting. The first is simple and well known thresholding where the measured position values are compared pairwise and thresholding violations registered. As dissimilar data noise strengths can make thresholdin unreliable the second method proposed by the authors is supplemented with an additional statistical evaluation where the measurements undergo a two-sample Z-test. Both methods were evaluated off-line with Monte-Carlo computer simulation. The tests showed that the proposed statistical method outperforms the straightforward thresholding approach.
