Skip to main content
Research
  1. Queen Mary University of London
  2. Research
  3. Research Strategy
  4. Research Culture at Queen Mary
  5. The CIRCLE Fund
  6. Safer Shared Roads

Safer Shared Roads: Modelling Cyclist Behaviour, Perceived Safety, and Environmental Exposure

Project team

Prof Isabelle Mareschal – Visual Cognition (SBBS)
Dr Mona Jaber – Reader in Internet of Things (EECS)

Project description

Early studies indicate that cyclists' accidents occur when they feel safe; as such, investigations that aim to improve cyclists' safety by merely modelling their safety perception are misleading. This research intends to jointly model the cyclists' behaviour and their safety perception in relation to road furniture, traffic scenes, and exposure to air quality. The outcomes will inform regulations and policies on how to make roads safer for cyclists who share the space with other vehicles and vulnerable road users.

How did the team come together?

The collaboration was formed after a faculty research away day, at which Isabelle had made a pitch about creating an initiative around the topic of ‘Healthy East London’. Mona approached Isabelle after this to discuss her work, leading to a CIRCLE project application as way to kick start the collaboration.

How did you decide on this question/topic?

Mona had been working on cycle safety for a while and had an early set of results, but she was finding it complex to model as she needed expertise to help predict the human behaviour of cyclists in reaction to risk. For example, how do they behave on the streets? How will different settings will trigger them to do something different e.g. being stuck behind a bus?

What activities will you undertake as part of this project?

Data will be collected from participants cycling an 8 km road from Stepney Green to Whitechapel via Bethnal Green and Old Street. Bicycles will be equipped with GoPro cameras, accelerometers, air-quality sensors, and GPS trackers. Participants will also wear smart watches and annotate moments where they feel safe/unsafe. This rich, multi-modal dataset will be fused with two external sources: a distributed acoustic sensor network providing continuous cyclist speed data along the same route, and historical records of accident locations.

The team will then develop a model that integrates these diverse data streams to understand how participants' behaviour, perception, and external factors jointly determine safety outcomes. Following this, they will analyse air quality variation by road type and quantify the damage of inhaled pollution as a function of cycling speed.

A student challenge will be organised based on the curated dataset and a workshop will be planned to showcase the findings.

Back to top