AllTheDocks road safety dataset: A cyclist's perspective and experience

TL;DR

AllTheDocks addresses the lack of open, ground-truth data on urban cycling safety and comfort by collecting a large-scale, citizen-science dataset in London. The authors combine helmet-mounted GoPro video with GPS, accelerometer, and gyroscope telemetry and obtain frame-level safety labels and hazard annotations, along with a road-roughness metric based on IRI and a Modified IRI. Key contributions include open access to video and sensor data, a systematic labeling workflow for safety perception, and methods to estimate road quality from sensor streams. The dataset aims to support urban planning and policy decisions to improve cycling safety and encourage mode shift from motorized transport to cycling.

Abstract

Active travel is an essential component in intelligent transportation systems. Cycling, as a form of active travel, shares the road space with motorised traffic which often affects the cyclists' safety and comfort and therefore peoples' propensity to uptake cycling instead of driving. This paper presents a unique dataset, collected by cyclists across London, that includes video footage, accelerometer, GPS, and gyroscope data. The dataset is then labelled by an independent group of London cyclists to rank the safety level of each frame and to identify objects in the cyclist's field of vision that might affect their experience. Furthermore, in this dataset, the quality of the road is measured by the international roughness index of the surface, which indicates the comfort of cycling on the road. The dataset will be made available for open access in the hope of motivating more research in this area to underpin the requirements for cyclists' safety and comfort and encourage more people to replace vehicle travel with cycling.

Figures (7)

• Figure 1: The accelerometer data fluctuates more than GPS speed. The maximum peak of 3D acceleration is caused by a sudden drop on the road, showing the strong influence of road roughness on acceleration.
• Figure 2: Trajectory covered; (Blue) routes that are represented with video footage and (Red) the route of timelapse images.
• Figure 3: Entire span of routes covered during the All The Docks challenge by 5 teams.
• Figure 4: Visualisation of telemetry data embedded in GoPro video footage using https://goprotelemetryextractor.com/free/
• Figure 5: When cyclists are stopping, it generates IRI peaks that are irrelevant to road roughness the same as most peaks in raw IRI (blue) have shown. (Note: in each time s, IRI calculation covers a road section at time s$\pm 2.5$ seconds since our IRI time window is set to 5 seconds.)