Aircraft Air Data Estimation Using Internal Avionics and Simulated Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional air data systems on aircraft rely on external sensors that are vulnerable to damage, create drag, and interfere with aerodynamics, making it challenging to determine accurate air data parameters, especially in high-performance military aircraft with limited resources and complex maneuvers.
Innovation Solution
A machine learning model using internal avionics data to predict air data parameters in real time, trained with simulated data from an integrated flight control model, wind tunnel testing, and computational fluid dynamics, without relying on external sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If external sensors are used to measure air data parameters, then measurement precision is improved, but reliability deteriorates due to vulnerability to damage and environmental factors
Solution Approach 1:
The patent creates a virtual copy of the external sensor system by training a machine learning model (neural network) to replicate the function of physical external air data sensors. The model is trained on simulated training data that mimics real sensor measurements, enabling it to predict air data parameters (airspeed, altitude, angle of attack, angle of sideslip) with accuracy comparable to external sensors while avoiding their vulnerability to damage and environmental interference.
Solution Approach 2:
The patent replaces the mechanical/physical sensor system with a computational/software-based system. Instead of using physical external sensors that are vulnerable to damage, the invention uses a machine learning model that processes internal avionics data to estimate air data parameters. This substitution eliminates the physical vulnerabilities while maintaining measurement capability through computational prediction.
2Measurement precision
If external sensors are installed on the aircraft, then air data measurement capability is improved, but object-generated harmful factors worsen due to drag and aerodynamic interference
Solution Approach 1:
The patent uses internal avionics data (which already exists in the aircraft system) to create a virtual measurement system that replicates the functionality of external sensors. By processing data from internal sources through a trained machine learning model, the system achieves air data measurement capability without installing additional external sensing hardware that would create drag and aerodynamic interference.
Solution Approach 2:
The invention replaces the physical external sensor installation with a computational approach using existing internal avionics. This substitution eliminates the need for external sensor housings, mounting structures, and associated aerodynamic interference, while still providing the required air data measurement capability through software-based estimation.
3Measurement precision
If more sensors are added to improve measurement accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent makes the machine learning model serve multiple functions: it processes data from various internal avionics sources (inertial measurement unit, flight control system, engine control unit) and predicts multiple air data parameters (airspeed, altitude, angle of attack, angle of sideslip) simultaneously. This multi-functionality allows the system to achieve comprehensive air data measurement accuracy without adding separate dedicated sensors for each parameter, thereby avoiding increased device complexity.
Data Source
AI summary
A computer-implemented method of computing training data for training an air data estimator (106) to predict values of air data parameters (120) of an aircraft (100) comprises: selecting ground truth values of the air data parameters (120) according to a flight envelope of the aircraft (100). The method involves simulating corresponding values from avionics in the aircraft (100) by using stored empirical data about engines of the aircraft (100) and empirical data about the aircraft (100) obtained from wind tunnel testing, and rules of computational fluid dynamics.