Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AsaPy: A Python Library for Aerospace Simulation Analysis

Joao P. A. Dantas, Samara R. Silva, Vitor C. F. Gomes, Andre N. Costa, Adrisson R. Samersla, Diego Geraldo, Marcos R. O. A. Maximo, Takashi Yoneyama

TL;DR

The paper presents AsaPy, a Python library that unifies established data-science techniques for aerospace simulation analysis without introducing new methods. It defines a modular architecture with four components (doe, execution control, analysis, prediction) wrapped around popular libraries, enabling end-to-end workflows from experimental design to ML-based prediction, and demonstrates the approach with military BVR case studies. Key contributions include automated hypothesis testing, distribution fitting, feature scoring, Pareto analysis, and a flexible ML pipeline built on TensorFlow and Scikit-Learn, applied to engagement decisions, WEZ prediction, missile hit-prediction, and navigation-fuel studies. AsaPy emphasizes interoperability with ASA and other simulators, supports batch execution with early stopping, and provides geodesy utilities, aiming to streamline aerospace data analysis and empower researchers with a scalable, open-source tool. The work discusses potential future enhancements, including expanded DoE and ML options, performance optimizations, and broader real-world demonstrations to validate efficiency gains.

Abstract

AsaPy is a custom-made Python library designed to simplify and optimize the analysis of aerospace simulation data. Instead of introducing new methodologies, it excels in combining various established techniques, creating a unified, specialized platform. It offers a range of features, including the design of experiment methods, statistical analysis techniques, machine learning algorithms, and data visualization tools. AsaPy's flexibility and customizability make it a viable solution for engineers and researchers who need to quickly gain insights into aerospace simulations. AsaPy is built on top of popular scientific computing libraries, ensuring high performance and scalability. In this work, we provide an overview of the key features and capabilities of AsaPy, followed by an exposition of its architecture and demonstrations of its effectiveness through some use cases applied in military operational simulations. We also evaluate how other simulation tools deal with data science, highlighting AsaPy's strengths and advantages. Finally, we discuss potential use cases and applications of AsaPy and outline future directions for the development and improvement of the library.

AsaPy: A Python Library for Aerospace Simulation Analysis

TL;DR

The paper presents AsaPy, a Python library that unifies established data-science techniques for aerospace simulation analysis without introducing new methods. It defines a modular architecture with four components (doe, execution control, analysis, prediction) wrapped around popular libraries, enabling end-to-end workflows from experimental design to ML-based prediction, and demonstrates the approach with military BVR case studies. Key contributions include automated hypothesis testing, distribution fitting, feature scoring, Pareto analysis, and a flexible ML pipeline built on TensorFlow and Scikit-Learn, applied to engagement decisions, WEZ prediction, missile hit-prediction, and navigation-fuel studies. AsaPy emphasizes interoperability with ASA and other simulators, supports batch execution with early stopping, and provides geodesy utilities, aiming to streamline aerospace data analysis and empower researchers with a scalable, open-source tool. The work discusses potential future enhancements, including expanded DoE and ML options, performance optimizations, and broader real-world demonstrations to validate efficiency gains.

Abstract

AsaPy is a custom-made Python library designed to simplify and optimize the analysis of aerospace simulation data. Instead of introducing new methodologies, it excels in combining various established techniques, creating a unified, specialized platform. It offers a range of features, including the design of experiment methods, statistical analysis techniques, machine learning algorithms, and data visualization tools. AsaPy's flexibility and customizability make it a viable solution for engineers and researchers who need to quickly gain insights into aerospace simulations. AsaPy is built on top of popular scientific computing libraries, ensuring high performance and scalability. In this work, we provide an overview of the key features and capabilities of AsaPy, followed by an exposition of its architecture and demonstrations of its effectiveness through some use cases applied in military operational simulations. We also evaluate how other simulation tools deal with data science, highlighting AsaPy's strengths and advantages. Finally, we discuss potential use cases and applications of AsaPy and outline future directions for the development and improvement of the library.
Paper Structure (18 sections, 5 figures, 1 table)

This paper contains 18 sections, 5 figures, 1 table.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. Existing Solutions
  4. Related Concepts
  5. STRUCTURE
  6. Design of Experiments
  7. Execution Control
  8. Analysis
  9. Prediction
  10. Support module
  11. CASE STUDY: BEYOND VISUAL RANGE AIR COMBAT SIMULATIONS
  12. Engagement Decision Support
  13. Weapon Engagement Zone Evaluation
  14. Missile Hit-Prediction
  15. Fighter Aircraft Navigation
  16. ...and 3 more sections

Figures (5)

  • Figure 1: The structure of the AsaPy library in the perspective of the analyst workflow.
  • Figure 2: Flow diagram for hypothesis testing using AsaPy.
  • Figure 3: Examples of simulation platforms in which AsaPy may be employed: FLAMES (left) and ASA (right).
  • Figure 4: Linear regression of Time of Flight vs. Fuel Consumed.
  • Figure 5: Comparative analysis of aircraft performance parameters. (a) presents a 2D perspective, while (b) offers a 3D visualization, providing a comprehensive overview of fuel efficiency dynamics.