CRSLP: Cooperative Roadside LiDAR Placement Benchmark with 900+ Plans

CRSLP is a benchmark dataset for evaluating cooperative roadside LiDAR placement strategies. It covers 13 sites, over 100 traffic scenarios, and more than 900 placement plans with corresponding long IoU sequences. The dataset, created by Yang Ma and last updated in 2026, includes over 3600 frames of point cloud data per test configuration, totaling 1.6 GB.

Use Cases

• Benchmarking sensor placement optimization algorithms based on the provided long IoU sequences.
• Simulating and evaluating traffic perception coverage across over 100 different traffic scenarios.
• Training or validating models for cooperative perception using the provided point cloud data segments.

Strengths

• Includes more than 900 distinct cooperative roadside LiDAR placement plans for evaluation.
• Covers a diverse set of 13 different sites and over 100 traffic scenarios.
• Provides at least 3600 frames of point cloud data per test configuration, with over 100 segments offered.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The large volume of point cloud data requires a manual request and download process.
• Row count for any potential tabular data is unknown, which may limit suitability assessment.

Provenance

Source

Yang Ma, via figshare.

Collection Method

Likely collected from simulated or real-world roadside sensor deployments across multiple sites.

Freshness

Last updated 2026-06-02 04:28:55; freshness should be verified.