Radio Ray Tracing Optimization via AutoDiff
Published:
Keywords: Python, JAX, Ionosphere, RadioRayTracing, Automatic Differentiation
This study advances the field of radio wave propagation modeling, particularly through the ionosphere, by integrating automatic differentiation (AD) with a stiff ordinary differential equation solver. This innovative method enhances the accuracy of radio ray trajectory predictions, based on their initial direction, and offers prospects of increased computational efficiency and precision. Despite the present challenges in refining this approach, it represents a substantial progression in ionospheric radio wave propagation modeling. The study is accompanied by Python code to practically illustrate the application of this concept. As for future endeavors, the focus will be on applying AD to discrete models for further optimization and problem-solving. Additionally, efforts will be dedicated to harnessing GPU accelerations for handling larger models, paving the way for even more efficient computational processes.