Aircraft Trajectory Prediction Using a Fuzzy Inference Framework via ADS-B Track Data

Flight trajectory prediction is crucial for aviation automation and management. Traditional methods struggle with nonlinear stochastic behaviors, while deep learning methods are computationally costly for long-term predictions. We propose a novel fuzzy inference framework that captures nonlinearities and utilizes actual data for interpretable modeling. Our framework also introduces a membership function for high-dimensional vectors, ensuring physical explainability. Compared to Transformer methods, our approach achieves the same accuracy while reducing processing time and memory usage.