STOP: enhanced safety on GCF and GEFER worksites

More machines equipped with AI optical sensors

GCF and GEFER reaffirm their commitment to safety on railway worksites by equipping additional locomotives with “STOP”, the innovative computer vision system designed to detect human presence in danger zones and alert operators, thus preventing accidents on site.

Following its experimental debut on the tracks about a year and a half ago, STOP – Supervised by Trained Optical Protection was installed last year on five locomotives operating on worksites, proving highly effective as a risk mitigation device capable of preventing hazardous situations and potential accidents. The new phase of the project, launched in recent weeks, now includes installation not only on seven additional locomotives but also on other types of operational machines used on worksites, primarily loaders, which have already successfully undergone testing in recent months.

Developed through the collaboration between the Research & Development Department of GCF S.p.A. and the Bologna - based company Kiwitron S.r.l., specialized in safety devices, STOP consists of one or more AI optical sensors connected to an onboard monitor and to visual and acoustic warning systems.

Thanks to months of engineering and field testing, the system -installed on operational machines typically used in railway track worksites - has been trained to operate both day and night, at working and cruising speeds, in all weather conditions and, above all, to recognize and distinguish human figures present in “danger zones” - even when seen from behind or bent over - and to report their presence to cab operators without confusion with “false positives” within the monitored area, such as objects, signage, or trackside materials.

One of STOP’s key advantages is that it does not require workers to use any additional signalling devices, such as battery-powered PPE, which could be cumbersome and intrusive.
Another important feature is that it is an AI optical system that complies with privacy regulations (GDPR), is equipped with internal computational capacity, and is pre-engineered to ensure compliance with EU AI Act requirements, particularly regarding non-self-evolving AI systems..

Its adoption, now extended to additional locomotives and further types of construction machinery, represents a significant strengthening of GCF’s corporate safety standards.

Video: how STOP works

Although still a relatively young project, STOP has already received significant recognition. In November 2024, it was presented at the 6th National Conference “Railway Safety and Operations” (SEF 24), held in Rome on 7 November at the Faculty of Civil and Industrial Engineering of Sapienza University of Rome. This prestigious event, organized by For.Fer - Training Centre recognized by the Italian National Agency for Railway and Road Infrastructure Safety (ANSFISA) and CIFI (Italian Association of Railway Engineers), brings together experts, companies, and institutions to explore technological innovations aimed at improving railway safety.

The system was developed within the broader OK (Zero Kill) plan, in which GCF S.p.A. is investing significant resources with the goal of creating safer and more secure worksites by preventing workplace accidents. Ongoing initiatives include Hansei, the Signalling System, and other innovative solutions currently under testing.

Thanks to STOP, GCF not only reduces the risk of accidents on site but also promotes an innovative approach to safety based on the integration of advanced technologies and responsibility.
The gradual extension of the system represents a concrete step toward increasingly safer worksites, protecting operators and strengthening corporate standards.

{ampz:ampzshortcode}