Challenge

STIB-MIVB (Brussels Intercommunal Transport Company) is the public transport operator in Brussels, Belgium. More than 1,300 vehicles (tram, bus, and metro) operate on the STIB-MIVB lines, moving an equivalent of the entire Brussels population on a daily basis. Performing efficient and robust monitoring of the STIB-MIVB tram infrastructure is of the utmost importance to ensure the smooth running of their operations. However, despite the STIB-MIVB maintenance efforts, trams break down on average once every 2 weeks due to issues with electrical lines, causing delays, traffic and many unhappy citizens of Brussels. To avoid these problematic situations, this project had several objectives. First, to anticipate and detect mechanical problems on overhead lines in near real-time. Secondly, to collaboratively develop and implement an efficient monitoring tool for the STIB-MIVB maintenance staff. Thirdly, to optimize the developed solution performance and resource consumption.

‍

Outcome

Our team collaborated extensively with STIB-MIVB and Bagaar to develop an innovative monitoring system based on a custom data processing and AI driven camera system. We developed an AI-driven image processing solution to detect key points on electrical lines and deliver pertinent information regarding the network state in near real time. The state-of-the-art tool has undergone meticulous optimisation to guarantee high-level accuracy while maintaining low resource consumption, striking a balance between performance and cost efficiency. Our developers conceived and implemented a visual platform, showcasing the gathered data directly onto a dynamic map of Brussels. The innovative solution allows for easy and efficient daily diagnostics, thus significantly improving the efficiency of network monitoring. Furthermore, to ensure seamless network maintenance, we integrated automatic alarms and alerts into the system. The feature is designed to detect any anomalies promptly, providing the opportunity to conduct preventive maintenance on trams and electric lines, and thereby mitigating the risk of major disruptions or malfunctions.

‍

Conclusion

Our AI-driven solution combined the development of a custom data management system and the use of AI on video data to provide near real-time monitoring of the STIB-MIVB tram network. The benefits of this project are manifold as it helps the network maintenance crew with their day-to-day operations and ensures the reliability of the network for its many daily users.