OpticalNeuromorphicEikonal_complex_512: Pathfinding Benchmark on a 512x512 Grid

262,144 grid cells represent a 2D pathfinding problem with 30% obstacle density and a random speed field. The dataset was created by Francisco Angulo de Lafuente to benchmark GPU-accelerated navigation algorithms, specifically the Optical Neuromorphic Eikonal Solver. The solver achieves a 135x average speedup over CPU Dijkstra with sub-1% accuracy.

Use Cases

• Benchmarking pathfinding algorithm speed and accuracy based on the described 512x512 grid with obstacles.
• Evaluating algorithm performance on variable-cost terrains based on the random speed field.
• Comparing neuromorphic or GPU-accelerated solvers against classical methods like Dijkstra based on the provided performance metrics.
• Testing navigation algorithms in structured, discretized environments based on the 2D grid representation.

Strengths

• Explicitly defined grid size (512x512) and obstacle density (30%) provide a controlled benchmark environment.
• Performance benchmarks for the associated solver are quantified with specific metrics (135x speedup, 0.64% mean error).
• Includes variable terrain costs via a random speed field, adding complexity for algorithm evaluation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Francisco Angulo de Lafuente via GitHub and HuggingFace Datasets.

Collection Method

Synthetically generated for algorithm benchmarking.

Time Range

null

Freshness

null

Geography

null