Freight Rail, Passenger Rail, Products & Technology, Digitalisation, Technology and IT, Safety, Signalling & Communications

HORUS AI improves driver awareness

Safe train operations require the train driver to be constantly aware of the locomotive’s precise location, potential hazards ahead of the train, the train’s speed relative to the permitted speed, all while maintaining compliance with the limit of authority or trackside signals. 

Combine reliance on constant driver situation awareness over a multi-hour shift with variables such as night operations, adverse weather, and the possibility of unauthorised access in the rail corridor, and the result is the likelihood of a safeworking incident.   

The typical risk control for freight train operations is the presence of the second driver, yet the number of safeworking incidents each year seems to indicate this control is vulnerable to human error.  The underlying cause of many of these incidents can be attributed to loss of situation awareness by the train crew. 

This then poses the question of how to leverage emerging technology to assist and improve train driver situation awareness and train protection. 

Newcastle-based company 4AI Systems seems to have found the answer with its development of HORUS – an artificial intelligence machine learning system that can be configured as an independent Advanced Driver Advisory System (ADAS) to improve driver situation awareness.  

4AI chief executive officer Joanne Wust tells Rail Express that by utilising complex neural networks to process data in real-time, the HORUS ADAS uses real-time processing of data from onboard sensors such as visual and thermal cameras, INS and GPS to detect and classify objects in the rail corridor such as people, large animals and vehicles.  

“It can detect hazards in the rail corridor at distances in excess of 1000 metres depending on the sensor configuration,” she said.  

“The HORUS ADAS can also monitor train compliance with a movement authority limit and permitted track speeds.   

“For signalled track, the HORUS ADAS can warn the train driver of the approaching signal and the displayed aspect.  Train braking curves can be calculated in real-time for train speed compliance with an approaching speed reduction or limit of authority. 

“Warning and alarms can be presented to the train driver.  ADAS can also be interfaced to the train brakes to brake the train should the driver fail to comply with alarms advice presented via the train driver display.” 

This technology has not gone unnoticed, with 4AI Systems receiving continued interest from operators for applying the HORUS ADAS to local and international rail networks.  

Presenting at InnoTRANS 2022 earlier this year in September, 4AI Systems showcased for the first time HORUS’ success in a remote and rugged environment – presenting footage from the Pilbara region of Western Australia.  

Visitors to the 4AI Systems stand were shown footage capturing accurate detection ranges of objects up to 1500 metres in front of a moving train, including signals, track workers, vehicles, and livestock across the remote rail network.   

4AI Systems also demonstrated HORUS’ detection and classification capability in North America’s largest metropolitan railway, in New York. 

The New York ‘proof of concept’ was a result of 4AI Systems being selected as one of four companies to participate in the New York City Transit Tech Lab Signalling Challenge to trial new technology to implement a more efficient and reliable subway signalling system more quickly and cost-effectively.   

As part of the proof, HORUS was installed on a R143 subway car along the Canarsie line to measure people, objects, and signal aspects, presenting an opportunity for Australian rail technology to be recognised in the international rail space. 

“The rail industry is fortunate to be able to leverage the car industry’s research and development in support of autonomous vehicles,” Wust said.  

“Innovative capabilities such as HORUS ADAS offer the opportunity to improve driver situation awareness and train protection and improve rail freight competitiveness against trucks.    

“The rail industry also has a significant advantage that unlike a car, the technology can be retrofitted to locomotives.”  

The company will showcase the HORUS AI technology at AusRAIL, at stands 105 and 110.