Keywords: smart railway maintenance, autonomous, artificial intelligence, maintenance automation, automated railway systems, smart operation and maintenance.

This research successfully demonstrated the feasibility of autonomous ultrasonic rail inspections up to technology readiness level (TRL) 7. A prototype integrating an autonomous rail vehicle and the latest artificial intelligence (AI) Sperry's ultrasound testing (UT) system was used for these demonstrations. The project initially demonstrated TRL5 attainment at Cranfield’s Railways Innovation Test Area (RITA). It was then prepared for a series of tests at a heritage operational railway to achieve TRL 7 attainment. Experimental works included nine rounds of tests on a 250-meter track inspection, showcasing inspection, localization, navigation accuracy, and defect location precision. The prototype successfully detected all the simulated rail defects and reported to the command centre as required. The vehicle performance was characterised by measuring its positional error and detection rate. The verified odometry and GPS positional measurements revealed errors ranging from 0.27 to 3.2 m and up to 8 m, respectively. The absence of differential GPS and an unrefined fusion approach contributing to these errors. Weak 4G signal coverage during field tests impacted operator-vehicle communication and data uploading rates. Future iterations should address these limitations, exploring alternatives for enhanced accuracy and advancing defect-sizing technology.

download the full-text of this paper (PDF, 1153 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description