COGNITIVE RAIL PILOT
An AI-based Assistance System for Engine Drivers
Accuracy, safety, and expedience are the key principles of efficient railroad transportation and logistics. Each transport system, however, is a complex organism with its own challenges.
Errors in operating locomotives can result in severe accidents and losses; heavy traffic or high-intensity sorting of railroad cars increases the risk of dispatcher errors and locomotive downtime, and the human factor may lead to violations of transportation rules and lower efficiency of rail transport.
The emergence of AI technologies is rapidly changing our world, including the rail transport industry. Cognitive Rail Pilot is an AI-based system designed to help engine drivers avoid dangerous errors that may lead to accidents, and to help locomotive fleet operators boost the efficiency of logistics management.
COGNITIVE RAIL PILOT MAKES TRANSPORTATION SAFE
Cognitive Rail Pilot will be able to identify a red signal or a closed switch.
Cognitive Rail Pilot will be able to identify a human or the danger of collision with another train.
Cognitive Rail Pilot will warn the driver, lower the speed of the locomotive, or perform braking if necessary.
COGNITIVE RAIL PILOT MAKES TRANSPORTATION EFFICIENT
Logistics management requires precise information on a locomotive’s whereabouts. By combining data from GPS, wheel rotation sensors, radar, and video cameras, our system knows where and on which track the locomotive is located with an accuracy of 30 cm, even at multi-track stations.
A high-precision map of the railroad section is an important component of the system. The suite charts it in Mapping Mode. Rail Pilot charts the tracks and key infrastructure elements (lights and switches) as well as contact-line poles, which the system uses as reference points in the primary operating mode.
HOW COGNITIVE RAIL PILOT SEES THE WORLD
COGNITIVE RAIL PILOT COMPONENTS
- Video camera unit with electronic climate control and maintenance system and three-axis vibration dampener
- Specialized high-resolution millimeter-wave radar
- High-performance industrial-grade computing unit
HOW COGNITIVE RAIL PILOT UNDERSTANDS ITS SURROUNDINGS
ORIGINAL DEEP-LEARNING NEURAL NETWORKS AND POST-PROCESSING ALGORITHMS
1
Distinguish between several main classes of railroad objects
2
Identify signals
3
Detect the position of switches
Cognitive Rail Pilot
demo video
demo video
On July 19, 2019, we presented Russia’s first, and also one of the world’s first, prototypes of a railroad locomotive with autonomous driving functionality. Cognitive Pilot is developing this project jointly with Russian Railways.
CALCULATE THE ROI OF COGNITIVE RAIL PILOT
Want To Learn More About Our Products?