Synthetic Flight Performance Data Generation Using GAN Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for generating synthetic flight performance data are limited by data quality issues and diversity, and real-world constraints restrict the number of actual flights that can be undertaken to gather sufficient data.
Innovation Solution
A method using a generative adversarial network (GAN) with a synthetic aircraft performance data generator and discriminator is trained to produce realistic and diverse synthetic flight data by learning patterns from actual aircraft performance data, improving the generator's ability to mimic real data and the discriminator's ability to distinguish between real and synthetic data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If actual flights are undertaken to gather flight performance data, then data volume increases, but time consumption and resource usage increase
Solution Approach 1:
The patent uses a generative adversarial network (GAN) to create synthetic copies of real flight performance data. The generator network learns the distribution patterns of actual flight data and generates synthetic data samples that statistically resemble real data, eliminating the need to conduct additional physical flights to increase data volume.
Solution Approach 2:
The patent replaces the mechanical process of conducting actual flights with an computational system. Instead of physically flying aircraft to collect data, a neural network-based system processes and generates synthetic flight data, substituting physical experimentation with information processing.
2Quantity of substance
If current systems generate synthetic data, then data volume increases, but data quality and diversity deteriorate
Solution Approach 1:
The patent implements a feedback mechanism through the discriminator network in the GAN framework. The discriminator evaluates synthetic data samples and provides feedback signals to the generator, guiding it to improve data quality and realism. This adversarial feedback loop continuously refines the synthetic data generation process.
Solution Approach 2:
The patent employs dynamic training where both the generator and discriminator networks are continuously updated during training. The system transitions from static data generation to a dynamic learning process where the generator adapts its output based on the discriminator's evaluations, progressively improving data quality.
Data Source
AI summary
Systems and methods of synthetic flight performance data generation include obtaining training data corresponding to multiple timeseries of multivariate aircraft performance data of actual aircraft flights. The systems and methods also include performing a training operation of a generative adversarial network that includes a synthetic aircraft performance data generator and a discriminator. After completion of the training operation, the systems and methods also include receiving one or more input parameters, and generating, at the synthetic aircraft performance data generator, one or more timeseries of synthetic aircraft performance data based on the one or more input parameters.