Airplane Flight Phase Identification Using Maximum Posterior Probability Method

The flight of a civil airplane is supported by many different systems. However, during the flight pilots require different navigation and surveillance data to perform airplane guidance. Each flight phase requires the usage of different methods of guidance and maintaining a preplanned trajectory. Automatic flight phase identification is an important component of on-board data processing and in the ground surveillance data processing unit. In the paper we study plight phase identification using a probability-based approach. Flight phase is detected by vertical speed and current barometrical altitude by method of the maximum posterior probability. A normal probability density function is used as a conditional function due to the assumption of normal error distribution of barometrical sensor. The proposed approach has been verified with real traffic surveillance data obtained under automatic dependent surveillance-broadcast communication channel within Ukrainian airspace. Classification on four common flight phases of taxing, take-off, en-route, and landing is used. The proposed method guarantees flight phase recognition with a minimum risk of false detection and indicates well detection in case of gaps in data sets.