The Artificial Intelligence Institute of Samara National Research University demonstrated its developments at the project session on applying for state support measures for integration of innovative technical solutions in railway transport. The project “End-to-end digital analytics platform based on machine learning, using an automated control system for unmanned aerial vehicles (ACS UAV)” was presented to Oleg Nikolaev, the Head of the Centre for Innovative Development of Russian Railways JSC, and the seniors of the Kuibyshev Railway.

According to Artem Nikonorov, Director of the Artificial Intelligence Institute, the system functionality consists of two modules. The first one is responsible for controlling unmanned aerial vehicles, which take photos and videos of infrastructure facilities, while inspecting them. The second module is engaged in photogrammetric processing of the received images, with subsequently using them for linear and volumetric measurements of the facilities. Simultaneously, the neural network based on deep machine learning identifies defects and damages to buildings and structures, making it possible to reach the decision on their repair or reconstruction.

The participants of the project session got interested in the development options, the effect of its application, as well as key technical solutions underlying it.

According to Artem Nikonorov, the comparison showed that the presented digital platform provided measurement error about 3 times less than that of common software products of the similar purpose – only about 3 % against 9–10 %. Moreover, the development of the Artificial Intelligence Institute has an additional opportunity – to provide hyperspectral survey of facilities using hyperspectrometers, also developed at Samara University and surpassing world analogues in technical and commercial specifications.
Automating the process of inspecting buildings and structures can significantly reduce the time spent, as well as make this work safer, since now there is no need for specialists to be present in person in hard-to-reach parts of infrastructure facilities, including those at altitude.

“The end-to-end digital platform has been developed since 2021, and the project has currently passed the testing phase. The system has already been tested at the facilities of the Kuibyshev Directorate for Traction Rolling Stock, the Kuibyshev Directorate for Track Repair. The Central Directorate for Track Repair is preparing to purchase a representative number of licenses, for providing regional divisions on the entire railway network”, Artem Nikonorov said.

For extensively using the end-to-end digital platform in the railway transport sector, it needs to be finalized taking into account the experience gained. This process can be accelerated by attracting financial support measures provided by the state for innovative projects aimed at solving the tasks of large industrial enterprises, including Russian Railways JSC.

The Ministry of Economic Development of Russia, together with the Analytical Centre under the Government of the Russian Federation, will have held by the end of 2023 the competition aimed at selecting new research centres based on leading universities for formation of artificial intelligence technologies, for the purpose of allocating appropriate grant support. Samara University has started preparing the application for participating this competition, as well as studying the opportunity for Russian Railways JSC to participate as a partner.