261,121 grid cells represent a 2D pathfinding problem with 30% obstacle density, created by Francisco Angulo de Lafuente. The dataset is designed to benchmark GPU-accelerated navigation algorithms, specifically the Optical Neuromorphic Eikonal Solver, which achieves a 135x average speedup over CPU Dijkstra. It includes columns for coordinates, obstacles, speed fields, and source/target points.
Use Cases
- Benchmarking pathfinding algorithm speed and accuracy based on the described 511x511 grid with obstacles.
- Evaluating navigation in variable-cost environments based on the speed field column.
- Testing neuromorphic or GPU-accelerated solvers against a CPU baseline based on the reported performance metrics.
- Studying the Eikonal equation for wavefront propagation in structured grids.
Strengths
- Defined grid size of 511x511 cells (261,121 total) with a specified obstacle density of 30%.
- Includes performance benchmarks for the associated solver, reporting a 135x average speedup and 0.64% mean absolute error.
- Clear column structure described for coordinates, obstacles, speed, and source/target points.
Limitations
- Row count is unknown, which may limit suitability assessment.
- Column-level documentation beyond the listed names is absent; field semantics must be inferred after download.
- Last update date is unknown; freshness unverified.
Provenance
- Source
- openml
- Collection Method
- Synthetically generated for benchmarking, using a specified random seed (11).
- Time Range
- Publication year referenced is 2025.
- Freshness
- Last updated: unknown
- Geography
- Spatial coverage: null