Drive Cycle Data from a 3D Urban Traffic Simulator (ABM) in Unity

Sedar Olmez's repository contains datasets from a 3D agent-based traffic simulator. Each vehicle type, PHEV and ICEV, was simulated across nine different scenarios varying traffic density and speed limit adherence. The output includes columns for agent positions, speeds, collisions, and vehicle dynamics parameters.

Use Cases

• Modeling vehicle energy consumption based on simulated drive cycles for PHEV and ICEV types.
• Analyzing traffic flow and collision patterns under different vehicle densities and driver compliance scenarios.
• Training reinforcement learning agents for autonomous driving using detailed vehicle state and sensor data like raycastLength.
• Studying the impact of vehicle parameters like mass and downforce on simulated driving behavior.

Strengths

• Data is generated from a controlled simulation with nine distinct scenarios covering three traffic densities (10, 50, 100 vehicles) and three levels of speed limit adherence.
• Includes detailed vehicle dynamics parameters such as tractionControl, VehicleMass, and Downforce, which are often absent in real-world traffic datasets.

Limitations

• Row count is unknown, which may limit suitability assessment for large-scale machine learning tasks.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is synthetic, generated from a specific Unity-based simulator, and may not perfectly reflect real-world driving conditions.

Provenance

Source

Sedar Olmez via paperswithcode.

Collection Method

Generated by a 3D Urban Traffic Simulator using Agent-Based Modeling (ABM) in Unity (version 1.1.0).