Qualifying round - Will the final evaluation randomize lighting, shadows, or sun position?

In a recent forum post for the AI for Industry Challenge, a participant raised a critical question about the qualifying round's final evaluation: whether lighting, shadows, or sun position will be randomized compared to the public toolkit. Specifically, the user asked if the organizers plan to vary the pose, intensity, or color of lights in the aic.sdf file, adjust which lights cast shadows or their softness, or add a directional sun with randomized azimuth and elevation. The public toolkit currently includes none of these variations, making the question essential for teams preparing their perception and navigation algorithms.

This query reflects a common challenge in simulation-based robotics competitions: ensuring that solutions generalize beyond the exact training environment. Randomizing environmental parameters such as lighting and shadows tests a system's robustness to real-world conditions, where lighting is rarely constant. The organizers' response will likely influence how participants design their AI pipelines, especially for vision-dependent tasks. Related discussions on leaderboard scores and hardware setups indicate a community eager for transparency to optimize their submissions.