Synthetic Data Generation for Robust Sensing & Perception

This project explores synthetic data generation for sensing and perception workflows, with a focus on controlled scene variation, repeatable simulation, and dataset inspection.

The work uses simulation to vary environmental conditions and stress perception behavior before deployment in real-world scenarios. By changing weather, visibility, and scene parameters in a controlled setup, the dataset provides a practical way to evaluate how robust a perception pipeline is when the operating conditions shift.

The supporting reports and presentation below document the dataset generation workflow, the perception motivation, and the dataset structure used to organize the exported simulation outputs.