Harsh Landing Ultimate Sensor Fusion

This research paper presents a sensor fusion methodology aimed at improving the accuracy and reliability of navigation systems in civil and general aviation, particularly during approach and landing. The objective is to provide sufficient accuracy required to perform automated landing which is a major goal for leading aviation companies like Airbus. The conventional INS/GPS solution has limitations due to weaknesses in GPS, exacerbated by challenging environmental conditions and increasing air traffic. The proposed methodology combines data from gyroscopes and accelerometers as inertial references, GPS as the primary observer, and Radio Altimeter (RA) and Instrument Landing System (ILS) as backup observer sensors when GPS is unreliable. An extended Kalman filter was developed and optimized using ground truth datasets to process the diverse sensor data. In addition, validation of the methodology was conducted using an X-plane plugin to simulate various landing scenarios at Montréal-Mirabel International Airport (CYMX) on runway 06. The results demonstrated improved positioning accuracy during the landing phase compared to the conventional INS/GPS solution, with a 50% enhancement in overall 3D positioning accuracy. The fusion approach offers several advantages over alternatives. It requires minimal hardware modifications to aircraft and airports, making it a cost-effective solution. Furthermore, it relies on radio avionic signals, reducing dependence on environmental conditions compared to vision-based solutions.