Ford Multi-Agent Seasonal Driving Data from Michigan, 2017-2018

2017-18 data collected by a fleet of manually driven Ford Fusion vehicles on a 66 km route in Michigan. The dataset captures seasonal variations in weather, lighting, construction, and traffic across urban, freeway, airport, and suburban scenarios. It was created by Ford Motor Company and includes raw sensor data, calibration, pose trajectories, and 3D maps in ROS bag format.

Use Cases

• Testing sensor fusion algorithms based on synchronized data from four 3D lidars and seven cameras.
• Developing robust object detection models based on seasonal variations in weather and lighting conditions.
• Simulating and validating multi-agent interactions based on data collected by a fleet of vehicles.
• Benchmarking localization and mapping (SLAM) systems against provided ground truth pose and 3D maps.
• Analyzing the impact of dynamic urban elements like construction and traffic on autonomous driving systems.

Strengths

• Captures seasonal variation across weather, lighting, and traffic conditions in a dynamic urban environment.
• Includes data from a suite of sensors: an IMU, four HDL-32E Velodyne lidars, and seven Point Grey cameras.
• Provides ground truth pose and 3D maps for validation and benchmarking.
• Data is collected on a defined 66 km route covering diverse scenarios like airports, freeways, and city centers.

Limitations

• Row count and dataset size are 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

Ford Motor Company

Collection Method

Collected by a fleet of manually driven Ford Fusion vehicles outfitted with IMU, lidar, and camera sensors.

Time Range

2017-2018

Freshness

null

Geography

Michigan, USA (specific route including Detroit Airport, freeways, city-centers, university campus, and suburban neighborhoods)