Seoul Robotics and Missouri-based Herzog Technologies have introduced an automated obstacle detection and warning system aimed at making the interface between the railroad and the public safer.

Seoul Robotics is a Korean 3D computer vision company using deep-learning artificial intelligence (AI) for mobility applications. Herzog is a railroad systems and technology integrator.

The Cam critical asset monitoring offering combines Seoul Robotics’ Sensr-I 3D perception software with Herzog’s occupancy detection platform with the goal of reducing future railroad incidents.

Cam uses a multi-sensor hardware platform and 3D perception engine to track and classify objects. This edge detection is coupled with business intelligence, which allows the system to analyse the information in real time and make timely decisions for notifying the appropriate entities. It can detect humans, vehicles, bicycles and other foreign objects, enabling it to provide safety insights.

A component of the system is the ability to evaluate the situation according to complex conditions as well as the behaviour of the objects being monitored. This enables the system to make smart decisions that are specific and unique to each area being monitored.

“We pride ourselves on equipping our partners with solutions that will improve safety and efficiency,” said Glen Dargy, vice president of technology at Herzog. “Cam provides customers with the ability to monitor aspects of their critical infrastructure in ways previously unavailable. With this innovative technology, our goal of improving safety is now within reach.”

Since 2011, over 23,000 incidents have occurred in the USA where trains have struck trains or people, resulting in 2700 fatalities and 9500 injuries, according to the US Department of Transportation. By using Cam systems at crossings, bridges, tunnels, stations and other areas of critical interest along the corridor, train conductors and operations personnel are programmatically alerted to hazards detected in monitored areas.

The technology is being implemented by Trinity Railway Express, a commuter rail operator, between Fort Worth and Dallas, Texas, with engineering underway for several additional bridges, crossings and passenger stations with other operators.

“For a rail detection system, every second is critical,” said Jerone Floor, vice president at Seoul Robotics. “Trains need adequate time to fully stop and require advanced warning to reduce the chance of a collision. This is only possible with an accurate, highly granular 3D perception system like ours. Through our partnership with Herzog, we are making this technology available to any rail operator, ensuring they can cost-effectively access these insights and make every second count.”

Sensr-I uses deep learning to analyse and understand 3D sensor data accurately. It enables the detection of more than 500 objects up to 200m ahead, predicts motion up to three seconds in advance, and provides real-time object perception.

The technology is compatible with more than 75 types and models of 3D sensors and supports systems consisting of different brands of sensors, making the technology cost-effective and scalable over large geographies. It also comes equipped with dynamic weather-filtering AI that enables Cam to glean and decipher obstacles even through harsh weather conditions, such as snow or heavy rain.

Founded in 2017, Seoul Robotics has partnered with OEMs, system integrators and government agencies around the world to diversify the use of 3D data. The company has developed proprietary software that is compatible with nearly all commercially available lidar and 3D data sensors to increase accuracy and efficiency and ensure safety across a range of industries and applications. The company has offices in Seoul, Silicon Valley, Munich and Detroit.

Herzog is an employee-owned company founded in 1969, headquartered in St Joseph, Missouri. With more than 2300 field and office personnel located across a network of construction and transit offices, Herzog operates in 36 states and maintains a large presence in California, Texas, Florida and Missouri.