IMUMETER-A Convolution Neural Network-Based Sensor for Measurement of Aircraft Ground Performance

Area of Science:

• Aerospace Engineering
• Artificial Intelligence
• Sensor Technology

Background:

• Accurate measurement of aircraft ground performance is crucial for aviation safety and efficiency.
• Existing methods for measuring aircraft ground distance during landing can be limited.
• The development of intelligent sensors is key to advancing aircraft motion analysis.

Purpose of the Study:

• To develop the IMUMETER sensor for precise aircraft ground performance measurement.
• To utilize artificial neural networks for automated aircraft motion analysis, specifically for landing distance.
• To create a sensor system integrating inertial and satellite navigation for enhanced data acquisition.

Main Methods:

• Development of the IMUMETER sensor based on artificial neural networks (ANNs).
• Integration of a single-board computer with an inertial navigation platform and satellite navigation sensor.
• Training a Convolutional Neural Network (CNN) using flight test data from a PZL 104 Wilga 35A aircraft.
• Testing and validation of original software solutions for sensor operation.

Main Results:

• Successful development and testing of the IMUMETER sensor.
• Accurate calculation of aircraft landing ground distance on a grass runway using the developed CNN model.
• Demonstration of the sensor's capability to recognize touchdown and stopping points for automated measurement.
• Exemplary measurements of landing distance provided for validation.

Conclusions:

• The IMUMETER sensor, powered by ANNs, effectively measures aircraft ground performance, particularly landing distance.
• The study validates the use of AI in developing advanced sensors for aircraft flight dynamics analysis.
• The developed technology offers a promising approach for future intelligent sensor systems in aviation.