Source: Fietsberaad, 23 March 2021. Image's source: Algemeen Dagblad

A trial is being conducted in Rotterdam with the live measurement of bicycle travel times and routes, without infringing on privacy.

Until now, apps or GPS tracking systems with a tracker on the bicycle have been used to measure travel times and travel routes. With such solutions you are depending on the cooperation of cyclists and privacy aspects are sometimes also at stake. Technolution developed Recall4bikes some time ago, which does not involve such problems. The system is now being deployed in Rotterdam.

Four boxes with cameras have been hung on the route from Central Station to the Erasmus Bridge, divided over two routes. The cameras are actually no more than image sensors, because a digital profile is recorded using a technique called deep learning of a cyclist, for example on the basis of the type of bicycle or the color of the jacket. The software converts the anonymized images into a match code. The match codes that are recognized with the next camera give an idea how much a certain route is used.

An extensive Privacy Impact Assessment was completed in Rotterdam for the study. This shows that the camera images are retained in the temporary memory of the camera system for a period of 0.3 seconds. This time is necessary to recognize cyclists and pedestrians and then convert them into a match code. This match code is then retained in temporary memory for 20 minutes. This period is necessary to enable re-identification at the next box.

Rotterdam hopes to learn from this trial what the favorite routes of cyclists are in order to be able to respond to this with bicycle policy.

In principle, the system can also be used - when travel times become longer - to adjust traffic regulations or to show alternative routes via a bicycle VMS.