Autonomous Train Control Overview

Autonomous Train Control (ATC) is an artificially intelligent system that adapts technologies utilized by autonomous vehicles to provide cost effective autonomous train operations. ATC employs a distributed and decentralized hybrid, customizable architecture which can be utilized in tandem with existing rail infrastructure or completely replace it through use of self-contained on-board hardware. Additionally, ATC can provide different levels of functionality, from simple safety enforcement to complete autonomous control.

Level 0

Leverages a powerful latent space odometry system designed for real-time sensing, along with deep sensor fusion, combining all the capabilities of all simultaneous cameras, GPS & IMU to achieve true redundancy. All of this makes possible deep scene understanding, trajectory mapping and cognition to make full operational decisions without physical control.

Level 1

Provides basic autonomous functionality to read aspects of signals & enforce maximum speeds, along with object & obstruction detection. Level 1 ATC utilizes computer vision to enforce maximum speeds during manual train operation and can be implemented within fixed-block, wayside and onboard signaling systems. An added benefit is that level 1 ATC can even be utilized in dark territories to provide a measure of protection against collisions.

Level 2

Allows for full autonomous operation. If used in conjunction with other signal technology, level 2 ATC can also be used to operate un-crewed vehicles. Level 2 ATC requires an onboard sensor suite for detection of vehicles ahead. This level of ATC provides the same level of vehicle detection as level 1, but it operates the vehicle as well. Level 2 can also be configured in a way to provide full train protection without the need for additional wayside signaling equipment.