The lifesaving tech saving lives at the subway

Content warning: Please be aware that this article talks about death by suicide and suicide prevention.

It is a tragic fact that many thousands of deaths occur each year where individuals put themselves in the path of a moving train. And rail networks all over the world are working to tackle an issue that is deeply distressing and traumatising for passengers, rail workers, train drivers and the families of those who heartbreakingly see this as their only option.

At one such network – a major U.S. metropolitan public transit organisation – police officers, control centre teams and operations managers are being supported by pioneers in video analytics technology, BriefCam – a Canon company. As you might expect, these subways and the passengers who use them are well served by a network of hundreds of cameras. However, the control rooms that monitor them are typically only staffed by one or two people and their capacity to spot every single instance of an individual moving towards or, oftentimes, lying down on the tracks is limited through no fault of their own.

Equally, it can be exceptionally difficult to assess a person’s intentions on a screen or for human eyes to identify someone who is just a little too close to the edge of a platform as a potential risk. A recent study of fatalities at stations found that those who attempted to end their lives in this way took a significant amount of time before they made their last move. This fact alone opens an important window of opportunity to save lives.

And the window that this transport organisation chose is polygonal.

The power of the polygon

Using the BriefCam solution, the team at the station can use their existing network of security cameras to pinpoint areas where a person is likely to head if their intention is to put themselves in the path of a train. Then the video analytics software allows them to place a virtual, adjustable polygon shape over the image of this access area, effectively creating a ‘geofence’ around it. A rule can then be configured in the software which triggers an alert if anyone crosses into the polygon and therefore is in an area of high risk.

It can even be adjusted only to respond to human figures moving into the space, so there are no false positives. In a particularly busy station, when customer-facing staff may find it difficult to spot those ‘red flag’ behaviours (such as lingering on the platform, pacing up and down, platform switching or moving to end of the track where the train approaches), having the high-risk spaces pre-identified and ringfenced can alert the control room of any breaches of these high-risk areas.

Of course, for the network to truly have a sense of whether this is an effective means to save lives would take the one thing happening that no one wants. And that’s precisely what happened when the system was being installed.

A BriefCam integration specialist was working with a station team, helping them to define their notification parameters, ruling out the movements of maintenance workers who routinely enter the space to undertake track servicing. They even added a rule to stop false alerts from exploring animals.

During the training session, an alert was triggered in the control centre – someone had crossed into the restricted area. The system operator immediately reviewed the alert and confirmed that a man was in the act of lying down on the tracks. This is the exact kind of pre-meditated behaviour that the BriefCam solution was installed to flag and so the station protocol for such a situation immediately began. An approaching train was stopped, police officers then raced to the scene and the man was helped to safety. The new solution had saved a life on day one.

Action, analytics and data-driven interventions

But besides this successful intervention and knowing that their new system achieved precisely what they needed, what did the team and police learn? Normally, they would need to try and piece together exactly what happened from witness statements and CCTV footage. In this case, BriefCam’s patented Video Analytics Platform (including VIDEO SYNOPSIS® technology) allowed them to quickly review, examine and understand the man’s behaviour, route and entrance to the tracks.

This then helped the team to put further proactive measures in place to reduce suicide attempts in their station. Because once you can see a potential pathway, you can interrupt it. The natural next step from here might be to start using BriefCam’s Artificial Intelligence powered tools to aggregate and analyse historical data from similar scenarios and attempts across the rail network. 

In this this implementation, video intelligence was intended for use by security and safety teams, but it could also oversee maintenance needs, be used for smart scheduling and apply pattern detection to help the network to understand the necessary staffing levels for their stations. Outside of the rail network, it can equally be scaled up for use in areas such as predicting traffic flow and crowd management and combined it’s these kinds of large-scale datasets that would be invaluable for use in urban planning. There is an argument that this is just the kind of connected ‘smart city thinking’ that will help us to understand civic spaces better and create a blueprint for cities that will keep us safer by design.

If you or someone you know is struggling with thoughts of suicide or needs immediate help, please contact emergency services or a local crisis helpline. You are not alone, and there are resources available to help.