Micro-level evaluation of bus stop performance using automated analysis of aerial video

Micro-level evaluation of bus stop performance using automated analysis of aerial video

Location, design and integration of public transport stops in the surrounding traffic environment are important factors affecting their micro-level efficiency, safety and comfort. Typical problems to name in this context are conflicts between entering/exiting passengers and cyclists passing the bus stop at high speeds, people taking shortcuts at illegal/unexpected places to avoid detours when reaching the stops, people running and making other risky movements when in a hurry to catch a bus and so on.

This is an issue of both working environment for the bus drivers for whom manoeuvring a large vehicle in a crowded/unpredictable environment could be very stressful, but also of safety and attractively of public busses as a transport mode in general. Additional 50–100 meters of illogical walking necessary to reach the bus stop could be a decisive factor in choosing between taking a bus or a private car (van Soest et al., 2019). Earlier studies (e.g. Berntman at el., 2012) have shown that while being on the bus, the passengers run very low risk of being injured in a traffic accident, yet while walking to/from the bus stop the risks are significantly (up to 100 times) higher.

The guidelines for designing the bus stops (e.g. Kol-TRAST) are usually quite verbal and give impression of being based on subjective opinions of what works well/less well. The capacity of the railway stations is studied sometimes with the help of microscopic modelling, yet this is seldom the case in the bus stop context. Moreover, the wider range of motion options and hidden preferences and (sometimes irrational) motivations of the bus users makes modelling of such environment much more difficult.

This project aims at adding objectivity in evaluation of bus stop performances by examining the microscopic behaviour of the passengers and other road users in vicinity of the bus stops. This will be done by the means of motion tracking in video taken from the unmanned aerial vehicle (UAV) which allows to cover a larger area (up to 100m) in vicinity of the bus stops necessary to properly understand the context of the motion activities and interactions.

Both UAV and automated video analysis technologies have seen a very rapid development on the technical side of things (Drone Centre Sweden, 2020; Data from Sky, 2020; Jensen et al., 2019). This project, however, will focus on searching for meaningful interpretations of the large amount of data from these tools that can aid the public transport professionals in evaluation and improvement of the bus stop designs and their integration in the surrounding environment.

Project manager: 
Aliaksei Laureshyn
Lunds universitet
350 000 kr
2020 to 2021