Reproducible planning study

Benchmark results

Sensitivity analyses and nominal-budget summaries for static buildings and fixed no-fly zones.

Descriptive benchmark · scene-clustered intervals

Loading recorded benchmark data…

Analysis scope

Sensitivity analyses

Raw-path excess is relative to the straight-line lower bound and is conditional on a successful, independently certified path.

Voxel resolution

Graph planners only. Markers are medians, vertical bars are interquartile ranges, and labels report successful runs over attempts. Missing estimates are not plotted.

RRT* sample budget

The horizontal axis is logarithmic. Markers are medians, vertical bars are interquartile ranges, and labels report successful runs over attempts.

Nominal-budget summary

Exactly one row per planner; no winner or rank is inferred.

Dataset-level estimates at the declared nominal budget. Time includes all runs; path and clearance metrics include successful runs only.
Planner Budget Path runs (scenes) Timing runs (scenes) Success % [95% CI] Time ms median [IQR] Raw excess % median [IQR] Smoothed excess % median [IQR] Clearance m median [IQR]

Download: run records (CSV), summary (CSV), dataset manifest, and timing provenance.

Method and provenance