Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal

Control

Sarah Salem

, Pascal Leone

and Axel Leonhardt

Institute of Transport and Spatial Planning, University of the Bundeswehr Munich, Munich, Bavaria, Germany

Keywords: Cycling Priority, Fuzzy Logic, Traffic Signal Control.

Abstract: The integration of cycling into urban traffic systems has increased significantly. Which drives the expansion

of dedicated bicycle lanes at intersections to accommodate the growing cyclist volumes while ensuring

traffic efficiency and safety. Addressing cyclists’ priority at signalized intersections presents a complex

challenge, necessitating tailored traffic signals and control methods. This research proposes a cycling

priority strategy for isolated intersections, using fuzzy logic to make high-quality decisions regarding cyclist

priority while minimizing delays for all road users. The methodology involves developing a fuzzy logic-

based cyclist priority strategy, using input variables such as vehicle queue and cyclist queue to determine

cyclist priority. The evaluation, conducted using VISSIM microscopic traffic simulation, demonstrates that

the proposed fuzzy logic-based control system effectively reduces delays and stops for cyclists, with an

optimal preference threshold (P*) value of 0.7 balancing the needs of both cyclists and motor vehicles.

Sensitivity analysis against traditional control methods further emphasises the potential of the fuzzy logic

approach to enhance overall traffic efficiency and promote sustainable urban mobility.

https://orcid.org/0009-0004-5581-9192

https://orcid.org/0009-0005-6265-1815

https://orcid.org/0009-0000-1382-3231

INTRODUCTION

Sustainable transportation is critical in metropolitan

areas to address pollution and traffic congestion.

Cycling has gained popularity due to its health

benefits, low emissions, and efficient use of road

space. However, integrating bicycles into traffic

systems, especially at intersections prioritising motor

vehicles, remains challenging. In Germany, cyclist

demand is particularly high, exceeding 800 cyclists

per hour per direction in Munich and 1,000 in Berlin

during peak summer hours (München, 2023;

Senatsverwaltung für Umwelt, 2023). These volumes

emphasise the need for improved cyclist

accommodation in traffic management systems and

infrastructure.

Studies show that stops and delays significantly

impact cycling experiences. Börjesson and Eliasson

(2012) revealed that cyclists perceive a one-minute

stop as equivalent to 3.1 minutes of cycling, reflecting

the greater effort and hazards associated with

interruptions (Börjesson and Eliasson, 2012). Fioreze

et al. (2019) found that cyclists often overestimate

waiting times by up to five times the actual duration

(Fioreze, 2019). Strategies like reducing signal cycle

lengths or extending green phases for cyclists are

cost-effective solutions to improve conditions, while

extensive infrastructure changes, such as segregating

bike and car flows, require higher investments (Gillis

et al., 2020; Poliziani et al., 2022).

The first section of this paper will review the

body of research on bike prioritising and traffic

signal regulation. The fuzzy logic-based (FL-based)

control system’s design and approach, including the

choice of input variables, membership functions, and

rule base, will next be presented. Subsequently, the

article will provide an overview of the simulation

environment and showcase the findings of research

that compares the suggested system with

conventional traffic signal control techniques. The

study will conclude with a discussion of the

findings’ implications and recommendations for

further research directions.

This study proposes a FL-based cyclist priority

strategy to address these issues. FL, a robust

490

Salem, S., Leone, P. and Leonhardt, A.

Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal Control.

DOI: 10.5220/0013292500003941

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 11th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2025), pages 490-497

ISBN: 978-989-758-745-0; ISSN: 2184-495X

artificial intelligence method for handling imprecise

data (Zadeh, 1975), is ideal for managing the

complexities of traffic flow. The system integrates

cyclist-specific factors like speed, acceleration, and

safety to improve travel experiences, reduce delays,

and promote sustainable transportation.

LITERATURE REVIEW

In recent years, the integration of bicycles into

traffic systems has gained attention, driven by the

growth of dedicated bicycle lanes at intersections

(Portilla et al., 2016; Wang et al., 2019). This

reflects efforts to accommodate increasing number

of cyclists while ensuring safe and efficient traffic

flow. Specialised traffic signals and control

strategies are essential for integrating bicycles

smoothly into intersections (Portilla et al., 2016).

For instance, Wang et al. (2019) proposed a group-

based signal timing model focusing on safety in

mixed traffic (Wang et al., 2019), while Portilla et

al. (2016) developed a predictive control system to

manage interactions between bicycles and vehicles

(Portilla et al., 2016).

FL has emerged as an effective tool for traffic

signal optimisation (Koukol et al., 2015; Pandey et

al., 2017). Introduced by Zadeh (1975), FL provides

a framework to manage uncertainties and imprecise

data in traffic systems (Zadeh, 1975). Studies have

used FL to prioritise specific road users, such as

emergency vehicles and public transit (Ikidid et al.,

2021; Chuo et al., 2022). Chuo et al. (2022)

demonstrated the use of FL for adaptive traffic

control, showing reduced delays and congestion by

dynamically adjusting signal timings based on queue

lengths (Chuo et al., 2022). Similarly, Nae and

Dumitrache (2019) applied FL to optimise signal

timings in urban intersections, significantly reducing

wait times and queues (Nae and Dumitrache, 2019).

Bhatia and Aggarwal (2020) highlighted the

environmental benefits of FL-based controllers and

suggested IoT integration for enhanced traffic

management (Bhatia and Aggarwal, 2020).

FL has also shown promise in transit and cyclist

prioritisation. Stevanovic and Teodorović (2022)

developed a Type-2 FL strategy to balance transit

and traffic delays, improving public transportation

operations while minimising disruptions to other

road users (Stevanovic and Teodorović, 2022). Vial

et al. (2023) explored cyclist prioritisation using

connected autonomous vehicles (CAVs) and noted

potential challenges, such as inconsistent

prioritisation and increased delays for cars (Vial et

al., 2023). Other approaches include using sensors

for cyclist priority during specific conditions, like

rain, or providing “green waves” via radar or mobile

apps (Fietsberaad, 2012; Verbeeke, 2020; Lai,

2021). However, cyclists’ low adoption of mobile

apps creates communication gaps, reducing their

effectiveness (Vial et al., 2023).

Integrating bicycles into urban traffic systems

through advanced strategies like FL is critical for

creating safer, more efficient traffic management.

These systems must balance the needs of all road

users, prioritising cyclists without significantly

disrupting motorised traffic (Gillis et al., 2020;

Poliziani et al., 2022).

METHODOLOGY

This research investigates a strategy to prioritise

cyclists at isolated intersections, adapting the

approach by Stevanovic and Teodorović (2022). The

primary objective is to develop a system using

approximate reasoning to make high-quality

decisions about cyclists’ priorities while minimising

delays for all road users.

Figure 1: Intersection Layout.

3.1

Fuzzy Logic Control

The system’s core component is FL, introduced by

Zadeh (1973). Fuzzy rules use descriptive

expressions like small, medium, or large to

categorise linguistic input and output variables,

creating a fuzzy control algorithm that quantifies

these expressions using fuzzy sets. This study

employs the fuzzy Mamdani logic method, also

known as the Max-Min method.

Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal Control

491

Figure 2: Membership Functions.

The cyclist priority strategy based on FL

includes two input variables—vehicle queue (VQ)

and cyclist queue (CQ)—and one output variable,

preference (P), which represents the percentage

preference to prioritise cyclists. VQ is the sum of

vehicle queue lengths in competing stages (VQ =

VQ1 + VQ2 + VQ3), as shown in Figure 1. At a

signalised intersection, vehicle/cyclist queue length

is defined as the distance from the stop line to the

rear of the last vehicle/cyclist waiting in any lane

during a red signal phase. Intelligent transportation

system technologies, such as smart roadside sensors

and advanced surveillance cameras, are being widely

adopted globally. These traffic cameras can function

independently or enhance the capabilities of other

smart roadside sensors, enabling the measurement

and detection of queue lengths in designated traffic

lanes with greater accuracy (Makino et al., 2018;

Umair et al., 2021).

The fuzzy sets for this system use triangle-

shaped membership functions, as shown in Figure 2,

to describe different categories. For vehicle queues

(VQ), there are three categories: Short, Medium, and

Long. These categories correspond to queue lengths

of 0 to 10 meters for Short, 10 to 50 meters for

Medium, and 50 to 100 meters for Long. Similarly,

for cyclist queues (CQ), the membership functions

define Short (0 to 5 meters), Medium (5 to 10

meters), and Long (10 to 20 meters). The output

variable, called “Preference,” is also divided into

three categories: Low, Medium, and High. This

variable determines how much priority cyclists

should get. The system uses an inference engine

with a set of rules to decide the level of preference

based on the input values for VQ and CQ. For

instance, if the vehicle queue is short but the cyclist

queue is long, the system gives high priority to

cyclists. On the other hand, if both queues are

medium, the system assigns a medium level of

priority to cyclists. The centroid method is used for

defuzzification, determining the crisp output value

by finding the “centre” of the area under the curve

formed by the membership functions.

Table 1: Fuzzy Rules.

Bike Q

Vehicle Q

Short Medium Long

Short Mediu

Low Low

Medium Hi

hMediu

Low

Long Hi

hHi

h Mediu

3.2

Cyclist Prioritization Strategy

Figure 3 shows a Pseudo code for the proposed

control strategy’s formulation through descriptive

rules. The decision on whether a cyclist approaching

the intersection should pass without stopping is

based on the detectors’ VQ and CQ values. A higher

CQ value increases the cyclist’s preference to pass

without stopping, while a higher VQ value decreases

this preference. The preference P (%) to prioritize

the cyclist can be low, medium, or high. If the FL-

calculated preference P exceeds a predetermined

threshold value P*, the cyclist should take priority.

The parameter P* value significantly impacts cyclist

and car delays. When cyclists are given priority,

actions are taken to allow them to pass through the

intersection without stopping. The cyclist signal

group extends its green light until the cyclist queue

clears or the maximum green time for that stage is

reached. If the cyclist signal group is showing a red

light, it transitions to a green light. This thorough

approach ensures a robust analysis and evaluation of

the proposed cyclist priority strategy, aiming to

improve the integration of cyclists into urban traffic

systems while maintaining overall traffic efficiency.

EXPERIMENTAL SETUP

We selected the three-legged intersection at the

entrance of the University of the Bundeswehr Test

track as the model area to assess the effectiveness

and robustness of our proposed methodology. For

simplicity we assumed that the cyclists are coming

from one direction. This setup allows us to

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

492

Figure 3: Pseudocode of FL-based Signal Control Strategy.

eventually test our algorithm on a real intersection.

We used the VISSIM microscopic traffic simulation

software by PTV AG, a tool commonly employed by

researchers in road traffic developments. VISSIM

features an intuitive graphical user interface (GUI)

for designing road network geometries and running

simulations. Additionally, the VISSIM-COM

interface creates a hierarchical

model enabling programmers to control simulator

functions and parameters initially set by the GUI.

Programmers can use any language that supports

COM objects, such as C++, Visual Basic, Java, or

Python.

Figure 4: Traffic Stages.

Signal data is managed by the VISSIG module,

with VAP defining control logic and VisVAP

offering a graphical interface. Static signal data is

stored in PUA files, while control logic is in VAP

files. By modifying VAP files, researchers can

evaluate and optimize signal control strategies,

adjusting parameterized stage lengths to improve

traffic flow and efficiency. This methodology allows

for a thorough evaluation of cyclist priority

strategies within urban traffic systems while

maintaining overall traffic efficiency. The control

logic encompasses three stages (Figure 4), with lane

widths between 2.75 and 3.50 meters and a vehicle

composition of 5% heavy goods vehicles (HGV) and

95% passenger cars, traveling at an average speed of

50 km/hr. Maximum green times are set at 30, 30,

and 40 seconds for stages 1, 2, and 3 respectively,

with minimum green times of 7, 4, and 10 seconds.

We utilize VISSIM to extract data on vehicle and

cyclist queues, as well as performance measures for

all road users.

The developed code implements a traffic signal

control algorithm that uses FL to prioritize traffic

flow based on real-time vehicle and cyclist queue

lengths. This program is designed to run every

second, ensuring timely traffic signal adjustments at

the intersection. It begins by defining several

constants necessary for its operation, including the

minimum and maximum green times for the three

traffic signal stages. These constants are parameters

for the FL subroutine and the main traffic control

logic, ensuring the program operates within set

limits. Figure 5 shows part of the signal control logic

implemented in the Vis-VAP module, with the FL

calculations executed using a Python script.

RESULTS AND ANALYSIS

This section evaluates the performance of the

proposed FL-based cyclist prioritization system

across various traffic scenarios. The results focus on

understanding the impact of the preference threshold

on key traffic metrics, such as delays and stops for

both cyclists and vehicles. The analysis also

compares the FL-based system’s performance

against traditional traffic control strategies to

highlight its relative advantages.

5.1

Validation Analysis

This section presents the results of our evaluation of

the proposed FL for a traffic signal control program

with varying preference threshold P* values. The

objective of this evaluation is to understand how

different P* settings impact the average delays and

stops experienced by vehicles and bicycles. By

analysing these results, we aim to identify optimal

Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal Control

493

Figure 5: Example of the Logic implemented in VisVap.

configurations that balance the needs of different

road users and improve overall traffic efficiency.

Figures 6(a) and 6(b) illustrate the average delays

and average number of stops experienced by

personal cars (PKW) and bicycles under various P*

values, respectively.

Figure 6: Impact of Different P* Values on (a) Average

Delay and (b)Average Number of stops for Personal cars,

and Cyclists.

In Figure 6(a), the average delay for personal cars

shows a sharp decline as the P* value increases, with

the highest delay observed at P* = 0. This delay

decreases significantly up to P* = 0.3, after which the

reduction becomes more gradual, indicating a

balancing trend. For bicycles, the delay exhibits the

opposite behaviour: it is lowest at P* = 0 and steadily

increases with higher P* values, peaking at P* = 1.

This trend highlights how prioritising bicycles (lower

P* values) effectively minimises their delays while

increasing delays for personal cars.

Figure 6(b) shows the average number of stops

for personal cars and bicycles. Similar to the delay

trends, the number of stops for personal cars

decreases as P* values increase. The highest number

of stops is observed at P* = 0, while fewer stops

occur as P* approaches 1. Conversely, cyclists

experience the fewest stops at P* = 0, with the

number of stops gradually increasing as P* values

rise, reaching a maximum at P* = 1. These patterns

emphasize the trade-off in optimizing delays and

stops for either personal cars or bicycles, depending

on the prioritization set by the P* value.

Based on the results, a preference threshold of

P*= 0.7 was selected as the optimal setting. This

value balances minimizing delays and stops for both

personal cars and cyclists. While lower P* values

(closer to 0) strongly prioritize bicycles, they result

in significantly higher delays and stops for personal

cars. Conversely, higher P* values (closer to 1)

disproportionately favor personal cars at the expense

of bicycle delays. The intermediate value of P* = 0.7

offers a compromise, reducing the gap between the

two road user groups and achieving a more equitable

and efficient traffic management solution.

Table 2: Traffic Scenarios Data.

Base 1 2 3 4

North-East 500 1000 500 500 500

West-East 600 600 600 600 1200

North-West 300 300 600 300 300

East 400 400 400 800 400

Cyclists 500 500 500 500 500

To validate the effectiveness of our proposed

logic and optimize its parameters, we compared the

average delay and stops produced by the fuzzy logic

(FL) controller (with P* = 0.7) to a fixed control

plan and a strategy from the German traffic light

guidelines (RiLSA). RiLSA is a technical standard

in Germany that includes specifications and

recommendations for planning and operating traffic

signals. The specific strategy used for comparison is

RiLSA Freigabezeitanpassung (FZA), which adjusts

the green time for each signal group based on

inbound gap time data. This strategy extends the

green time for a signal group with detected demand

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

494

Table 3: Comparison of Average Delay and Person Delay.

Avg Delay Vehicles Avg Delay Cyclists Person Delay

Fixed RiLSA FZA Fixed RiLSA FZA Fixed RiLSA FZA

Base 0.40 0.05 -0.63 -0.49 -0.13 -0.15

1 0.78 -0.21 -0.60 -0.34 0.07 -0.24

2 0.89 0.10 -0.63 -0.58 0.14 -0.13

3 1.46 0.07 -0.62 -0.46 0.43 -0.05

4 0.62 0.55 -0.66 -0.53 0.04 0.15

Figure 7: The Effect of Cyclist Demand on Various Traffic Performance Metrics. (a) Avg Delay for Vehicles, (b) Avg

Delay for Cyclists, (c) Avg person Delay including Pedesterians, (d) Avg no of Stops for Vehicles, (e) Avg no of Stops for

Cyclists.

(in this case, cyclists) within pre-defined minimum

and maximum limits, without altering the stage

sequence. To comprehensively evaluate

performance, we employed metrics such as average

vehicle delay, total cyclist delay, and overall person

delay, assuming an average vehicle occupancy rate

of 1.5 passengers per vehicle. Table 2 outlines the

different demand scenarios used in this evaluation.

In this analysis, negative values indicate that the

FL controller produces less delay compared to the

baseline strategies, while positive values indicate an

increase in delay. Our results demonstrate that the

proposed logic significantly outperforms both the

RiLSA and fixed-time controllers in reducing the

average delay for cyclists across all demand

scenarios. However, this improvement for cyclists

comes at the cost of a slight increase in the average

delay for personal cars. These results suggest that

the FL controller effectively prioritizes cyclists,

reducing their delays even under varying traffic

volumes. As shown in Table 3, the proposed logic

not only significantly reduces delays for cyclists but

also improves overall person delay. While delays for

personal cars increase slightly, the pronounced

benefits for cyclists result in a more favorable

overall performance. This highlights the ability of

the proposed traffic signal control program to

balance the trade-offs between different road users,

optimizing delays for cyclists without neglecting the

needs of vehicles and pedestrians.

5.2

Sensitivity Analysis Under

Different Cyclists Demand

The analysis conducted here aims to evaluate the

effect of cyclist demand on various traffic

performance metrics while maintaining vehicle

demand fixed. The comparison is made across three

strategies: FL with P* = 0.7, Fixed signal control, and

RiLSA FZA.

Figure 7 illustrates the performance of the FL

system across various traffic metrics under different

cyclist demand levels and in comparison with

Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal Control

495

traditional traffic control systems. For average delay

of vehicles (Figure 7a), the FL system shows a mode-

rate increase in delays as cyclist demand rises but

consistently maintains lower delays compared to the

fixed signal control. However, RiLSA FZA performs

better at higher cyclist demands, likely due to its lon-

ger green times allocated to clear queues efficiently.

In terms of average delay for cyclists (Figure 7b),

the FL system results in slightly higher delays

compared to RiLSA FZA, especially under higher

cyclist demand scenarios. This can be attributed to the

shorter green times allocated under FL, which aim to

balance traffic flow for both vehicles and cyclists. The

average person delay (Figure 7c), which includes

pedestrians, reflects a key trade-off between the three

strategies. While RiLSA FZA generally achieves the

lowest person delays, the FL system strikes a balance,

avoiding excessively high delays for vehicles. The

fixed signal control, in contrast, shows the highest

overall person delay, highlighting its limitations in

handling mixed traffic efficiently.

The average number of stops for vehicles (Figure

7d) shows that the FL system performs better than the

fixed signal control but is slightly less efficient than

RiLSA FZA at reducing stops as cyclist demand

increases, which indicates that FL provides smoother

vehicle flow. Finally, for the average number of stops

for cyclists (Figure 7e), the FL system shows

moderate performance, with fewer stops than the

fixed signal control but slightly more than RiLSA

FZA at higher cyclist demands. This is consistent with

the FL system’s balanced approach, which prioritizes

equitable green time distribution across all road users.

In summary, the fuzzy logic system demonstrates

a well-balanced approach to managing traffic at

intersections, effectively distributing green time

between vehicles, cyclists, and pedestrians. While it

may not outperform RiLSA FZA in cyclist-centric

scenarios, it provides a more equitable solution,

maintaining lower vehicle delays and fewer stops for

all road users compared to the fixed signal control.

As shown in Figure 8, the FL system activated

the bicycle signal less often than the RiLSA FZA

strategy, and the green time per actuation was

noticeably shorter. As cyclist demand increased, the

green time allocated per cyclist under FL stayed

fairly limited, ensuring that many cyclists could

cross during the green phase without excessively

prolonging the signal. This highlights how FL

prioritizes the efficient use of green time, enabling

cyclists to clear the intersection quickly, though this

comes at the cost of longer delays for larger cyclist

queues. In contrast, RiLSA FZA, which triggered

the bicycle signal more frequently, provided

significantly longer green times as demand grew.

This allowed it to clear larger queues in one cycle,

but it also sometimes resulted in unused green time

once all cyclists had passed. As a result, the green

time per cyclist under RiLSA FZA was higher

compared to FL. The shorter green times under FL

reflect its focus on balancing the needs of vehicles

and cyclists. By avoiding overly long green phases

for cyclists, FL minimizes excessive delays for

vehicles while maintaining fairness for all road

users. Although this approach might result in

slightly higher cyclist delays during periods of heavy

demand, it supports a more balanced distribution of

green time across the entire traffic system.

Figure 8: Effect of Cyclists Demand on Average Green

Time.

CONCLUSIONS

This study explored and tested a new way to prioritize

cyclists at traffic signals using fuzzy logic, a method

designed to handle the complexities of real-world

traffic. The system adjusts the signal timings dynami-

cally, taking into account the number of cyclists and

vehicles waiting at an intersection. The results indica-

te that this approach significantly reduces delays and

stops for cyclists, providing them with a smoother and

more efficient experience, while ensuring that vehicle

delays remain at acceptable levels. By maintaining

this balance, the system promotes a more equitable

traffic flow for all road users.

A key finding of the study was that a preference

threshold (P*) value of 0.7 worked best, ensuring

that cyclists and vehicles shared green time

equitably. While this approach slightly increased

delays for vehicles, it delivered noticeable

improvements for cyclists and reduced overall

delays for all road users. Compared to traditional

traffic control methods, this system stands out as a

practical solution to prioritize cyclists without

causing major disruptions to vehicle traffic.

The study also highlights the potential of fuzzy

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

496

logic to manage the unpredictable and ever-changing

nature of traffic. Its ability to adapt in real-time

makes it a significant improvement over fixed or

rigid systems. Beyond just benefiting cyclists, this

strategy supports larger sustainability goals,

encouraging more people to cycle by making it a

more attractive alternative to driving. This, in turn,

could help reduce emissions and contribute to

healthier urban environments.

Looking ahead, there’s room to refine this

system further. Future work could expand its design

to better include pedestrians and adapt to

intersections of different layouts. While this study

found that a P* value of 0.7 worked well, future

research could explore ways to make this value

adjustable in real time, optimizing performance

based on changing traffic conditions. The next step

will involve testing the system in a real-world

setting at the entrance of the University of the

Bundeswehr Test Track. These real-life trials will

help determine how effective and practical the

system is outside of simulation, paving the way for

broader adoption in urban traffic systems.

ACKNOWLEDGEMENTS

This research is part of the project MORE – Munich

Mobility Research Campus (MORE, 2023). The

project is funded by dtec.bw – Digitalization and

Technology Research Center of the Bundeswehr.

dtec.bw is funded by the European Union –

NextGenerationEU.

REFERENCES

Bhatia, M. S. and Aggarwal, A. (2020). Congestion

control by reducing wait time at the traffic junction

using fuzzy logic controller. International Journal of

Sensors, Wireless Communications and Control,

10(6):989–1000.

Börjesson, M. and Eliasson, J. (2012). The value of time

and external benefits in bicycle appraisal.

Transportation Research Part A: Policy and Practice,

46(4):673– 683.

Chuo, H. S. E., Seah, Y. E., Tan, M. K., Lim, K. G., Liau,

C. F., and Teo, K. T. K. (2022). On-demand priority traffic

optimizer with fuzzy logic microcontroller. In 2022

IEEE International Conference on Artificial Intelli-

gence in Engineering and Technology (IICAIET).

Fietsberaad (2012). Groningen gives cyclist the green light

more often. CROW.

Fioreze, T. (2019). Perceived waiting time versus actual

waiting time: a case study among cyclists in Enschede,

The Netherlands.

Gillis, D., Gautama, S., Van Gheluwe, C., Semanjski, I.,

Lopez, A. J., and Lauwers, D. (2020). Measuring

delays for bicycles at signalized intersections using

smartphone GPS tracking data. ISPRS International

Journal of Geo-Information, 9(3):174.

Ikidid, A., Fazziki, A. E., and Sadgal, M. (2021). A fuzzy

logic supported multi-agent system for urban traffic

and priority link control. Journal of Universal

Computer Science, 27(10):1026–1045.

Koukol, M., ZajíIková, L., Marek, L., and Tuček, P.

(2015). Fuzzy logic in traffic engineering: A review on

signal control. Mathematical Problems in

Engineering, 2015:1–14.

Lai, M. (2021). Intersection control for cyclists with

isignum.

Makino, H., Tamada, K., Sakai, K., and Kamijo, S.

(2018). Solutions for urban traffic issues by its

technologies. IATSS Research, 42(2):49–60.

München, L. (2023). Daten zum radverkehr.

Nae, A. C. and Dumitrache, I. (2019). Fuzzy-logic

adaptive control of traffic in an urban junction. U.P.B.

Sci. Bull., 81(2).

Pandey, S., Mathur, P., and Patil, T. (2017). Real time

traffic signal control using fuzzy logic controller:

Review. In Proceedings of the International

Conference on Innovations in Information, Embedded

and Communication Systems (ICIIECS).

Poliziani, C., Rupi, F., Schweizer, J., Saracco, M., and

Capuano, D. (2022). Cyclist’s waiting time estimation

at intersections, a case study with GPS traces from

bologna. In Transportation Research Procedia,

volume 62, pages 325–332.

Portilla, C., Valencia, F., Espinosa, J., Núñez, A., and De

Schutter, B. (2016). Model-based predictive control

for bicycling in urban intersections. Transportation

Research Part C: Emerging Technologies, 70:27– 41.

Senatsverwaltung für Umwelt, Mobilität, V.-u. K. (2023).

Fahrradverkehr in zahlen.

Stevanovic, A. and Teodorović, D. (2022). Type-2 fuzzy

logic based transit priority strategy. Expert Systems

With Applications, 187:115875.

Umair, M., Farooq, M. U., Raza, R. H., Chen, Q., and

Abdulhai, B. (2021). Efficient video-based vehicle

queue length estimation using computer vision and

deep learning for an urban traffic scenario. Processes,

9(10):1786.

Verbeeke, R. (2020). Green wave-apps for cyclists.

Vial, A., Salomons, M., Daamen, W., Van Arem, B.,

Hoogendoorn-Lanser, S., and Hoogendoorn, S. (2023).

Prioritizing cyclists at signalized intersections using

observations from connected autonomous vehicles.

Transportation Research Record: Journal of the

Transportation Research Board, 2677(12):29–43.

Wang, F., Tang, K., Li, K., Liu, Z., and Zhu, L. (2019). A

group-based signal timing optimization model

considering safety for signalized intersections with

mixed traffic flows. Journal of Advanced

Transportation, pages 1–13.

Zadeh, L. A. (1975). The concept of a linguistic variable

and its application to approximate reasoning—i.

Information Sciences, 8(3):199–249.

Real-Time Cyclist Prioritization with Fuzzy Logic-Based Signal Control

497