WINGSPARK: An anomaly detection tool for efficient monitoring, incidents detection and effects mitigation to enhance the security in railway infrastructure

May 24, 2022 | News

WINGS ICT Solutions ( designed and developed an extension of the WINGSPARK platform in order to provide the capability of detecting anomalous values in time series data and CCTV live feeds to show alerts and also provide potential mitigation measures (e.g., evacuation of a station in case of an incident etc.). WINGSPARK works together with the other tools in the SAFETY4RAILS context to provide a holistic approach in enhancing safety and security in trans-modal metro and railway networks. The alerts produced by the tool are sent using the SAFETY4RAILS Information System platform (S4RIS) Distributed Messaging System (DMS) to the RAM2 which in turn is responsible for their visualisation. WINGSPARK is also integrated in S4RIS GUI providing useful insights to the stations’ operators, enabling the efficient infrastructure monitoring and timely detection of potential incidents.

WINGSPARK is a platform which provides active monitoring, forecasting and anomaly detection mechanisms, effects mitigation while delivering insights regarding the operational condition of the infrastructure it supervises. Besides real-time monitoring, WINGSPARK provides anomaly detection components to detect potential incidents in the railway infrastructure that could result in potential cascading effects. For example, WINGSPARK platform takes as input the energy supply of a station and decides whether it could affect the speed of the train or not. If the energy supply is very low (energy deficit), there might be not enough power for the normal or appropriate operation of the train. Another component of the tool takes as input train speed data coming from IoT devices (sensors) in real time and triggers specified alerts when an abnormal behavior is detected based on a scoring mechanism. This is particularly useful in detecting potential failures in the signalling system or terrorist attacks that may cause a train to stop, which could in turn result in multiple train collisions and a magnitude of casualties.


Furthermore, WINGSPARK can detect overcrowded situations in the monitored railway infrastructure, based on video acquired through CCTV cameras. With this input, dynamic evacuation plans can be proposed in case of emergency.

More specifically, the evacuation service developed by WINGS aims at complementing existing processes and capabilities by providing automated guidance and respective evacuation routes to visitors and travelers, who are not aware of the facility. Real time data such as number of evacuees within an area, nearest exit, distance from the exit and visualization of the proposed evacuation route is provided.


The solutions regarding train spend and overcrowded areas have been showcased during the SAFETY4RAILS Metro de Madrid simulation exercise (organized in February 2022) and will also be demonstrated during the Rome and Milan exercises which will take place on 30 May/ 1 June and late June 2022 respectively.

In the future, the developed anomaly detection mechanisms could be utilized in the context of a risk assessment tool, which would provide the probability of anomalies detected evolving into a cascading effect. WINGSPARK has been configured to be able to accommodate such a pipeline, which could output the probability of a hazardous incident, providing valuable information and insights, to take measures to avoid such a situation.


Share This