Crowding is one the most common problems for public transportation systems worldwide. It has been proven to cause anxiety to commuters and create reliability problems when commuters are not able to board on the first train or bus that arrives. These commuters are referred as left-behind passengers, and their number is directly related to various basic performance measures of public transportation systems that represent the user’s experience. Among these measures the most significant are ridership, service quality and, more importantly, travel time. Identifying left behind passengers is a tool to address crowding in stations and respond appropriately, by applying various operational strategies such as decreasing headways.
The methodology proposed in this study has been applied to two stations with high probability of left behind passengers, Sullivan Square and North Station on the MBTA Orange Line in Boston, Massachusetts. Two types of technologies were used to detect passengers being left behind in the platform. The first one was an object detection software, namely You Only Look Once (YOLO), using surveillance cameras. The second type was a Bluetooth and Wi-Fi sensor mounted on the two selected stations. Moreover, manual counts of left behind passengers were collected in the two stations. Both technologies will be individually compared with the manual counts to test accuracy and precision. Finally, the two technologies are compared with the manual counts to determine a best way to detect left behind passengers.