Prediction of Hazardous Winds Directly Above a Vertiport for Urban Air Mobility Using a Pressure Observation Network

The concept of Urban Air Mobility (UAM), involving the operation of various small aircraft in urban areas, is gaining significant attention. In order to facilitate takeoffs and landings, access to vertiports on building rooftops and other locations is anticipated. However, encountering hazardous turbulence unique to urban spaces during the descent phase over vertiports could potentially lead to serious accidents. Therefore, the development of a warning system that provides both real-time and predictive turbulence information is imperative. To tackle this issue, this study proposes a method for predicting turbulence near the vertiport ground surface by deploying small pressure sensors in the surrounding buildings. A machine learning algorithm for predicting air turbulence based on multi-point pressure sensing is constructed and its effectiveness is shown through outdoor experiments. We will also propose a method to prevent significant degradation in prediction performance caused by pressure sensor failures.