Intersection control for cyclists with iSignum

Intersection control for cyclists with iSignum

Lai, Madeline (TU Delft Civil Engineering and Geosciences)

Salomons, A.M. (mentor)
Schakel, W.J. (mentor)
Boer, Deodaat (mentor)
Broekhuijsen, Peter (mentor)

Delft University of Technology

Civil Engineering

2021-10-18

The Netherlands is like no other country enthusiastic about cycling and is the home of 37.000 kilometers of bicycle paths. These bicycle paths need to cross other modalities and therefore intersections are needed. At signalized intersections, the control is mostly regulated using sensors such as push buttons and one or multiple inductive loops. Sensors for cyclists are running slightly behind in comparison with those from cars and the company of CycleData anticipated on this fact by developing a new sensor, the iSignum, that works on radar and laser service and is capable to determine speed. The iSignum is capable of communicating with the iVRI (intelligent traffic light control) where it is installed and notify that a cyclist is coming. The sensor is placed further away from the stop line meaning there is time to communicate with the iVRI that a cyclist was detected and for example with what speed. This can benefit the comfort and traffic flow of cyclists when a notification leads to green light such that cyclists simply do not have to stop unnecessarily before the intersection. Part of this research was a literature study towards stakeholders and existing bicycle sensors. The latter was needed to be able to make a comparison of what makes the iSignum stand out from the rest or if it is just another bicycle sensor. A literature study has indicated that inductive loop sensors in combination with push buttons are used the most. Additionally, there are other cyclists sensors on the market nowadays. These sensors include infrared sensors, smart cameras and mobile phones and WiFi/Bluetooth applications. These are however not used on a big scale yet. In this report, research was done towards the newly developed sensor that is installed at an intersection in Delft, the Julianalaan-Nassaulaan intersection. Several matters were looked into, such as the accuracy of the iSignum in comparison to the reality at this location and what potential situations can result in disturbances in detecting cyclists. Furthermore, preventions are explored to an extent of what could be possible at this specific location. To test the exactness of the iSignum, manual bike counting is done over the course of three days. This data was then compared to the data that the sensor of iSignum logs. Conclusions that have been made after the comparison is that the newly developed sensor is not as accurate as the company CycleData claims it to be, the promised 95% is not reached. During the measurements, the sensor did not manage to pass the 78.2% accuracy. However, disturbances that happened, such as the presence of a delivery van or garbage truck or cars that drive over the bicycle path, while counting cyclists are included. Making the preciseness less in particular cases, with the worst case being 75.3%. V-log data, which logs traffic status and information such as detection and signal groups from an iVRI, is additionally used to check whether those disturbances can be covered by the existing control of the intersection. The outcome states that the existing intersection control of a loop sensor and push-button manage to still detect cyclists during the time of disturbances. Next to this, prevention methods for the disturbances are researched such that relying on the existing intersection control may not be needed in the future and the full potential of the iSignum can be reached after more precision from the sensor itself is achieved. A prevention method that could help all the occurred disturbances is to move the sensor more downstream from the bicycle path where the disturbances do not influence the sensor anymore. This is therefore also the recommendation for this particular intersection such that unnecessarily errors are filtered out of the data. Additionally, further research is needed after the relocation of the sensor and towards the enhancement of the accuracy because other errors are present that are not accountable to a disturbance. It is valuable to do more extensive research on this topic to determine if these errors can be erased from the system to reach the full potential of the iSignum such that the objective of shorter waiting times and more comfort for cyclists can be reached at signalized intersections.

http://resolver.tudelft.nl/uuid:b29cd450-cefd-4f38-a5eb-04ccf4da78f4

Student theses

bachelor thesis

© 2021 Madeline Lai