E!79 ATTUNE, An intelligent predictive fault identification system for the railway industry, Irisity AB

Purpose and goal: Operations and Maintenance (O&M) activities in the rail sector are currently carried out during highly disruptive, manually intensive, expensive and infrequent periodic inspections. Dangerous faults can occur between inspections. ATTUNE will develop a disruptive, cost-effective rail infrastructure predictive monitoring platform to

reduce failure rates and improve network efficiency. Expected results and effects:

The project will develop low-cost sensor modules to generate data which will be processed on a cloud-based platform using machine learning. The sensor data, gathered during normal rolling stock operation, is interrogated using intelligent algorithms to identify anomalies and accurately predict when maintenance is required.

Crucially, the combination of visual and Inertial Measurement Unit (IMU) data from the three subsystems significantly enhances the predictive capability. The correlation of factors can better identify emergent faults and correctly identify root causes. Approach and implementation:

The project will be carried out by an international consortium. The main focus of Irisity will be computer vision related modules. Through data sharing between modules, the robustness of the total system is expected to increase beyond what is possible developing each module separately. E.g. data from the IMU may support visual odometry. Synergy effects with other research and development operations are expected.