Ford Autonomous Vehicle Dataset

2020: Opening of a data set

In March, 2020 Ford released in open access a data set, the unmanned vehicles assembled from the park. This array which received the name Ford Autonomous Vehicle Dataset will allow researchers to increase reliability of operation of self-governed machines in the urban environment.

For formation of a data set engineers of Ford reeled up kilometers on the hybrid Ford Fusion Hybrid sedans equipped with quad-core processors Intel Core i7 and 16 Gbytes of RAM on the route in Detroit. Robomobiles moved along the set routes, but slightly deviated them to cover "various" feature set.

Ford released a data set for the accelerated development of pilotless cars

Data were collected generally using four lidars (they measure distance to the purpose by its lighting by laser source and measurements of an indirect light), six 1.3-megapixel cameras, one 5-megapixel camera and the inertial measuring block.

On the basis of indications of sensors researchers of Ford generated cards and these provisions of pedestrians for sending to the Ford Autonomous Vehicle Dataset set. Data were registered in different weather conditions, including in sunny, snow and cloud weather and also include information on driving on highways, overpasses, the airports, bridges, tunnels, construction zones, etc.

According to developers, the data set contains primary data from sensors with a binding at the right time, calibration values, trajectories of objects in real situations and 3D maps. The data set is available in the file ROS format that allows to visualize data, to change and use them using the open ROS operating system (Robot Operating System).

This data set can form a basis for improvement of the modern robotic algorithms connected with multi-agent autonomous systems and make them steadier against seasonal and city changes — said in Ford.[1]

Robotics

Notes