Pedestrian and cyclist safety in the context of micromobility on the shared-use paths

A shared-use path is a route used by pedestrians, cyclists, and others with small non-motor vehicles. Across countries, different attitudes and preferences toward shared-use paths lead to varying challenges, with the main problem being limiting conflicts. In the Czech Republic, shared-use paths are increasingly used by e-bikes and scooters. Higher speeds and traffic volumes on shared-use paths have increased the number and severity of injuries. This research was set out to investigate the impact of measurable parameters on conflict prediction. Locations on 10 different shared-use paths were chosen, and 40 h of bird’s-eye video footage were recorded, analysed using the YOLO algorithm for detection of shared-use path users, and tracking their routes using the BoT-SORT algorithm. For all trajectories, lengths, average speeds, direction of motion, and time were calculated, along with mutual interactions. For every location, path width was measured, and conflicts and potential conflicts were evaluated. Volumes were processed into 15-minute intervals, and baseline characteristics were calculated. Machine Learning methods were used to analyse the dataset and visualise the results to better understand the conflict prediction model. The results confirm that higher traffic volumes increase the proportion of conflicts, while greater path width helps mitigate these risks. Conflicts are more likely with a higher proportion of cyclists and at lower speeds and higher volumes, suggesting congestion is a contributing factor. These findings underscore the importance of well-designed infrastructure, particularly wider paths, in enhancing safety and fostering harmonious coexistence among diverse user groups on shared-use facilities.