Optimising passenger flows through stations

**Background**

Railway companies have introduced zone marking on some station platforms to assist passenger boarding; although it is not clear that the way this is currently implemented is either understood by passengers or proving effective. Consequently, passengers are frequently not distributed optimally along the platform; and this results in congestion at train doors and longer than necessary boarding times as embarking passengers need to wait for disembarking passengers to leave the carriage before they can board. This issue is compounded when trains are delayed, especially at busy interchange stations. Our proposed solution will visually guide embarking passengers to demarcated zones on the platform based on the number of passengers leaving each carriage, to minimise the overall train boarding time and avoid “herd behaviour”. This will help to reduce train delays, minimise congestion and passenger discomfort, and improve platform safety by avoiding high passenger densities.

**Solution**

The proposed solution includes:

1. A method to detect passenger flows on the station platform using existing infrastructure including sensor and other data. The method will depend on linear algebra to deduce passenger flows that are not directly monitored.
2. A dashboard visualising current passenger flows and congestion levels. This will significantly increase the visibility of passenger congestion on the platform for operational station staff and enabling them to take informed decisions e.g. where to dispatch staff, where to nudge passengers; and for passengers to know where to move on the platform.
3. Development of an origin/destination matrix of passenger flows on the platform (i.e. disembarking and embarking passenger of arriving trains) from PFM RF beacons historical data in conjunction with open train arrival and occupancy data.
4. A method to calculate positioning of passengers along the platform, such that train boarding operations is optimised. The method is based on the predicted number of passengers embarking and disembarking. This will provide dynamic recommendations to passengers and station staff to facilitate a redistribution of passengers along the platform.

**Differentiation**

We believe that our proposed solution approach will convey three key advantages over those currently available:

1. 3D simulated CCTV images as an alternative real CCTV data which could be used to model scenarios and extreme events without compromising data privacy or having access to CCTV footage of extreme events.
2. We aim to use algorithmic approaches to infer flows across sensor "blind-spots"; thereby minimising the need for investment in significant new sensor infrastructure.
3. Generating semi-real passenger volumes using RF beacons installed at the station in conjunction with open train arrival and occupancy data.
4. Novell user interface to “nudge” passengers avoiding herd behaviour to exacerbate situation.

**Commercialisation**

Opportunities exist to scale the proposed solution across the UK and global rail industries and into other transportation sectors, including airports, that are facing similar challenges in facilitating the safe and smooth flow of passengers.