Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Exploratory Assessment of LLM's Potential Toward Flight Trajectory Reconstruction Analysis

Qilei Zhang, John H. Mott

TL;DR

The paper investigates whether large language models can reconstruct flight trajectories from irregular ADS-B data, focusing on the LLaMA 2 model with LoRA-based fine-tuning on synthetic, 60-second flight sequences. It shows that a base model struggles, but a fine-tuned model can accurately reconstruct both linear and curved trajectories, effectively filtering noise and handling missing ADS-B data within a 60-second window. However, token-length limitations prevent handling longer sequences and challenge real-time deployment, indicating that LLMs are not yet ready to replace Kalman filtering in precision-critical contexts. The work demonstrates the potential of LLMs for time-series processing in aviation and lays groundwork for future cross-model comparisons and expanded applications with longer sequences and richer evaluation metrics.

Abstract

Large Language Models (LLMs) hold transformative potential in aviation, particularly in reconstructing flight trajectories. This paper investigates this potential, grounded in the notion that LLMs excel at processing sequential data and deciphering complex data structures. Utilizing the LLaMA 2 model, a pre-trained open-source LLM, the study focuses on reconstructing flight trajectories using Automatic Dependent Surveillance-Broadcast (ADS-B) data with irregularities inherent in real-world scenarios. The findings demonstrate the model's proficiency in filtering noise and estimating both linear and curved flight trajectories. However, the analysis also reveals challenges in managing longer data sequences, which may be attributed to the token length limitations of LLM models. The study's insights underscore the promise of LLMs in flight trajectory reconstruction and open new avenues for their broader application across the aviation and transportation sectors.

An Exploratory Assessment of LLM's Potential Toward Flight Trajectory Reconstruction Analysis

TL;DR

The paper investigates whether large language models can reconstruct flight trajectories from irregular ADS-B data, focusing on the LLaMA 2 model with LoRA-based fine-tuning on synthetic, 60-second flight sequences. It shows that a base model struggles, but a fine-tuned model can accurately reconstruct both linear and curved trajectories, effectively filtering noise and handling missing ADS-B data within a 60-second window. However, token-length limitations prevent handling longer sequences and challenge real-time deployment, indicating that LLMs are not yet ready to replace Kalman filtering in precision-critical contexts. The work demonstrates the potential of LLMs for time-series processing in aviation and lays groundwork for future cross-model comparisons and expanded applications with longer sequences and richer evaluation metrics.

Abstract

Large Language Models (LLMs) hold transformative potential in aviation, particularly in reconstructing flight trajectories. This paper investigates this potential, grounded in the notion that LLMs excel at processing sequential data and deciphering complex data structures. Utilizing the LLaMA 2 model, a pre-trained open-source LLM, the study focuses on reconstructing flight trajectories using Automatic Dependent Surveillance-Broadcast (ADS-B) data with irregularities inherent in real-world scenarios. The findings demonstrate the model's proficiency in filtering noise and estimating both linear and curved flight trajectories. However, the analysis also reveals challenges in managing longer data sequences, which may be attributed to the token length limitations of LLM models. The study's insights underscore the promise of LLMs in flight trajectory reconstruction and open new avenues for their broader application across the aviation and transportation sectors.
Paper Structure (17 sections, 3 figures, 1 table)

This paper contains 17 sections, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Flight Trajectory Reconstruction
  4. LLM on Time-Series Data
  5. Methodology
  6. Flight Data Collection and Generation
  7. LLM Model Configuration
  8. Data Preprocessing
  9. Results
  10. Base Model Evaluation
  11. Fine-tuning Model Evaluation
  12. Linear Flight Trajectories
  13. Curved Flight Trajectories
  14. Discussion
  15. Limitations
  16. ...and 2 more sections

Figures (3)

  • Figure 1: An illustration of a simulated flight trajectory with synthetic ADS-B data departing from KLAF and arriving at KVPZ.
  • Figure 2: Flight trajectory reconstruction results for linear flight trajectories.
  • Figure 3: Curved flight trajectory with level and changing altitude.