OpticalNeuromorphicEikonal_sparse_128: Pathfinding Benchmark on a 128x128 Grid

A tabular dataset of 16,384 grid cells representing a 2D pathfinding problem with obstacles and uniform speed. It serves as a benchmark for the Optical Neuromorphic Eikonal Solver, which demonstrates significant speedups over CPU Dijkstra. The dataset was created by Francisco Angulo de Lafuente and is licensed under CC BY 4.0.

Use Cases

• Benchmarking pathfinding algorithm performance based on grid coordinates and obstacle flags.
• Evaluating navigation speed in uniform cost fields based on the speed column.
• Testing algorithm accuracy against near-optimal paths based on the reported 1.025x optimal length metric.
• Comparing computational efficiency against the reported 30-300x speedup over CPU Dijkstra.

Strengths

• Defined grid size of 128x128 cells, providing a structured test environment.
• Specific obstacle density of 10% and uniform speed field, enabling controlled comparisons.
• Performance benchmarks for the associated solver are provided, including a 135x average speedup and 0.64% mean absolute error.

Limitations

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

Provenance

Source

OpenML, with associated GitHub repository and HuggingFace Datasets page.

Collection Method

Synthetically generated for benchmarking purposes.

Time Range

null

Freshness

null

Geography

null