Electrical signal analysis
The need
Modernizing line and infrastructure management practices on a rail network requires a substantial increase in data digitization. The traditional preventive maintenance methods traditionally employed by operators are now proving to be unviable in the long term, given labor requirements. In the case of turnouts, the implementation of a predictive maintenance plan relies on the ability to carry out real-time analysis of the electrical transients generated when a switch is tilted from one track to another.
Our value proposition
Using our expertise in signal processing and the implementation of innovative neural network structures such as wavelet transforms and autoencoders, we have successfully demonstrated the validity of our mathematical hypotheses through a proof of concept (POC). Our ultimate aim is to develop a demonstrator prototype in future phases.
