Establishment and Verification of a Linear Regression Model for QAR Radio Altitude Correction

A landform can induce an abnormal change in the radio altitude of an aircraft during low-altitude approach. To address this problem, multiple linear regression (MLR) is used to investigate how various factors, such as pitch, roll and weight, affect the radio altitude and establish an altitude correction model for a low-altitude aircraft. A comparative analysis of the model prediction with the altitude above sea level based on local station pressure (QNH) indicates that the model can reflect the change trend in the radio altitude and correct the abnormal jump in the radio altitude change caused by a landform. Radio-altitude correction is useful for analysing warnings triggered by an abnormal radio altitude, eliminating false warnings, and improve the flight efficiency and safety margin.