Using Aircraft Onboard Surveillance and Navigation Signals for Unusual Aviation Weather Detection

Abstract

Existing unusual aviation weather detection systems use the weather information provided by several ground-based meteorological observation stations in the vicinity of airports. However, most ground meteorological observation stations are near the ground surface level, so they can only detect unusual aviation weather conditions through a proximate two-dimensional analysis. With only two-dimensional weather information, it is difficult to issue a warning for unusual low-level weather conditions, such as low-level wind shear and microbursts. To augment the existing system to provide weather anomaly information about 1,500 feet above the runway, the onboard surveillance and navigation signals broadcast by commercial aircraft via the automatic dependent surveillance broadcast (ADS-B) datalink is considered a source of vertical weather information. Thus, this paper is aimed toward determining whether aircrafts are encountering unusual aviation weather condition by using these data to acquire the weather information. To achieve this object, the correlation between actual aircraft onboard sensor data and unusual aviation weather conditions are evaluated based on several unusual aviation weather detection algorithms and machine learning method. Moreover, the verification and performance results and analysis of the proposed unusual weather detection method are discussed. The machine learning model can provide accuracy rates of close to 96% for detecting unusual aviation weather conditions. Finally, a graphic user interface is constructed to integrate these functions and then present unusual aviation weather information.