The award ceremony for the tenth edition of the ‘DITECFER Railway Innovation Contest’ took place on 3rd July in Pistoia as part of the DITECFER Ten-Year Anniversary Gala-Event Dinner, held at ‘La Cattedrale’ (formerly Breda), in front of 260 representatives of national industry and institutions.
The three winners were awarded by Prof. Alessandro Fantechi, President of the Jury of the Contest, Member of the Scientific Steering Group of Europe’s Rail, and President of the School of Engineering of the University of Florence. The Jury was composed, as every year, of representatives from Research and the Railway Industry.

The ceremony, full of emotion, saw the application of Artificial Intelligence (AI) as the trait d’union between the three winning innovations.
The top two prize-winners were united by their focus on the autonomy of trams: the centre of attention for increasingly innovative urban mobility.
Also the ‘DITECFER Railway Innovation Contest’ is celebrating its tenth anniversary this year: 30 ‘made in Italy’ innovations have been discovered and enhanced by DITECFER with this initiative over the years.
(Prof. Fantechi with the DITECFER Coordinator,
Veronica Elena Bocci, open the Award Ceremony)
“Autonomous Tram AI Technologies” is the innovation by Hitachi Rail GTS Italia (former Ground Transportation Systems Italia, former Thales Italia) that integrates Artificial Intelligence techniques with the innovative core functions NGAP (Next Generation Autonomous Positioning) and ODT (Obstacle Detection and Tracking). NGAP is GTS’s new positioning system that allows the tram to locate itself along the track line with an accuracy of less than 1 metre, in all environmental conditions, by combining measurements from IMU, radar and GNSS sensors through an advanced sensor fusion algorithm (SFA). ODT uses radar, lidar, cameras and AI to detect, classify and track obstacles around the tram. Sensor data is combined in a Multi Target Tracking (MTT) data fusion algorithm in order to detect obstacles with an extremely low rate of false negatives and false positives. These applications are containerised in a high-performance computer certified for railways. The applications are monitored in real time via a 5G connection. Together, NGAP and ADAS represent the key functions of GTS Italia’s innovative tram system.
The innovation has been tested for two years on trams in Florence.
(Prof. Fantechi awards the GTS team that developed the innovation:
from left to right, Eng. Luigi Rucher and Eng. Gianluca Mandò)

“HORA” (Hitachi autOnomous stReet cAr), the innovation developed by Hitachi Rail STS, provides an answer to the problem of tram drivers, who have to simultaneously monitor several critical areas (front area, blind spots, doors being (un)boarded, sides of the vehicle, etc.) and, as a result, may miss some dangerous situations. In addition, there are studies that have identified human error as the main cause of tram accidents, due to the pressure exerted by adherence to timetables, anticipation of other road users’ behaviour, tram operating constraints and fatigue due to workload.
The HORA system monitors the environment around the vehicle to detect, classify and track objects in critical areas, alerting drivers in real time to dangerous situations and suggesting the safest driving action, with particular attention to the safety of passengers and pedestrians.
(Prof. Fantechi awards the Hitachi Rail STS team that developed HORA:
from left to right, Eng. Gaetano D’Andria, Eng. Francesco Fasano, Eng. Luigi Fratelli
and Eng. Massimiliano Lenardi – the latter from Hitachi Rail Europe)

“MIRET-Tunnel AI” is the innovation by ETS classified third. It is the MIRET methodology software based on Artificial Intelligence (AI) algorithms for the detection of defects (to date defects due to water, detachments and cracks) of cladding structures (to date concrete/reinforced concrete, masonry, mechanised). MIRET mainly analyses transport infrastructures, which can be inspected with the ARCHITA acquisition system and are considered critical infrastructures. MIRET-Tunnel AI respects the fundamental criteria imposed by the AI Act, such as: data quality, security and robustness, transparency and information, documentation and human supervision. The algorithms used support expert users in the pre-assessment of the qualitative risk of the tunnel structures, following a quality design flow, which includes the generation of the vulnerability map, its verification, approval and validation, and a subsequent phase of dataset selection for training and enhancement of the AI models used. Personal data are protected as the information is clustered according to the architecture of the AI algorithms and can no longer be traced back to the source tunnel for the end user. MIRET-Tunnel AI is not a sold SW, but part of a risk analysis service in which many technical figures and responsibilities are involved.
(Prof. Fantechi awards Dr. Fabio Treglia, who receives the award
on behalf of the ETS R&D team led by Eng. Federico Foria)
