2962-780X

No URL ">
  •  
  •  
 

Abstract

Traffic management at at-grade pedestrian crossing facilities (pelican crossings) in highly populated areas, such as the Universitas Indonesia Station, faces significant inefficiency challenges. During peak hours, the fixed-time system is frequently disabled and replaced with subjective manual control by security personnel, thereby triggering irregular stop-and-go cycles and a high accumulation of vehicle delays. This study aims to develop a hybrid adaptive control model integrating Computer Vision, Genetic Algorithm (GA), and Fuzzy Logic to optimize intersection performance under mixed traffic conditions. The research methodology begins with the extraction of traffic and pedestrian characteristic data, calculated manually through recorded field observations. This manual extraction approach is conducted to ensure the accuracy of the baseline data (ground truth) and to avoid machine detection errors during the initial modeling phase. The precise empirical data is then used to calibrate the mathematical analytical model. Subsequently, a Genetic Algorithm is employed offline to determine the optimal delay equilibrium weights between vehicles and pedestrians. This equilibrium foundation is then integrated with Fuzzy Logic, which acts as a dynamic green time extension mechanism for the Computer Vision-based smart system architecture. Macroscopic evaluation results demonstrate that this hybrid adaptive model successfully performs cycle consolidation, effectively reducing the total number of signal cycles from 19 to 10 compared to the existing manual control. During the morning peak session, the system reduced cumulative vehicular delay by 13.03%, successfully saving a total of 737 seconds. This increase in vehicular efficiency necessitated a minor operational trade-off, resulting in a 5.17% increase in average pedestrian waiting time (from 15.29 to 16.06 seconds per person). Crucially, this adjustment remains strictly within the 18-second ideal safety tolerance threshold, proving that the model successfully achieves an ideal operational equilibrium between competing traffic demands.

Bahasa Abstract

Manajemen lalu lintas pada fasilitas penyeberangan sebidang (pelican crossing) di area padat populasi, seperti Stasiun Universitas Indonesia, menghadapi tantangan inefisiensi yang signifikan. Pada jam puncak, sistem waktu tetap (fixed-time) sering dinonaktifkan dan diganti dengan kendali manual yang subjektif oleh petugas keamanan, yang pada akhirnya memicu siklus berhenti-dan-jalan (stop-and-go) yang tidak teratur serta tingginya akumulasi tundaan kendaraan. Penelitian ini bertujuan untuk mengembangkan model kendali adaptif hibrida yang mengintegrasikan Visi Komputer (Computer Vision), Algoritma Genetika (GA), dan Logika Fuzzy untuk mengoptimalkan kinerja simpang dalam kondisi lalu lintas campuran. Metodologi penelitian diawali dengan ekstraksi data karakteristik lalu lintas dan pejalan kaki, yang dihitung secara manual melalui rekaman observasi lapangan. Pendekatan ekstraksi manual ini dilakukan untuk memastikan akurasi data dasar (ground truth) dan menghindari kesalahan deteksi mesin pada fase awal pemodelan. Data empiris yang presisi tersebut kemudian digunakan untuk mengkalibrasi model analitik matematis. Selanjutnya, Algoritma Genetika diterapkan secara luring (offline) untuk menentukan bobot ekuilibrium tundaan yang optimal antara kendaraan bermotor dan pejalan kaki. Fondasi ekuilibrium ini kemudian diintegrasikan dengan Logika Fuzzy, yang bertindak sebagai mekanisme perpanjangan waktu hijau dinamis untuk arsitektur sistem cerdas berbasis Visi Komputer. Hasil evaluasi makroskopis menunjukkan bahwa model adaptif hibrida ini berhasil melakukan konsolidasi siklus, secara efektif mengurangi jumlah total siklus sinyal dari 19 menjadi 10 siklus dibandingkan dengan kendali manual eksisting. Selama sesi jam puncak pagi, sistem ini mampu mengurangi tundaan kumulatif kendaraan bermotor sebesar 13,03%, dan berhasil menghemat waktu sebanyak 737 detik. Peningkatan efisiensi kendaraan ini membutuhkan kompromi operasional (trade-off), yang mengakibatkan peningkatan rata-rata waktu tunggu pejalan kaki sebesar 5,17% (dari 15,29 menjadi 16,06 detik per orang). Secara krusial, penyesuaian ini tetap terjaga ketat di dalam ambang batas toleransi keselamatan ideal yaitu 18 detik, membuktikan bahwa model tersebut berhasil mencapai titik ekuilibrium operasional yang ideal di antara pergerakan lalu lintas yang saling bersaing.

References

Abood, A. N. (2024). Traffic Light Control in Vehicular Network Systems using Fuzzy Logic. European Journal of Information Technologies and Computer Science, 4. https://doi.org/10.24018/ej.2024.4.4.132

AlKheder, S., & AlRukaibi, F. (2020). Enhancing pedestrian safety, walkability and traffic flow with fuzzy logic. Science of the Total Environment, 701. https://doi.org/10.1016/j.scitotenv.2019.134454

Cakici, Z., & Aksoy, G. (2024). Does the minimization of the average vehicle delay and the minimization of the average number of stops mean the same at the signalized intersections? International Journal of Transportation Science and Technology, 13, 213–228. https://doi.org/10.1016/j.ijtst.2024.01.003

Fahrunnisa, Z., Rahmadwati, R., & Setyawan, R. A. (2024). Adaptive Traffic Light Signal Control Using Fuzzy Logic Based on Real-Time Vehicle Detection from Video Surveillance. Jurnal Ilmiah Teknik Elektro Komputer Dan Informatika, 10(2), 235–251. https://doi.org/10.26555/jiteki.v10i2.28712

Hawas, Y. E. (2011). An integrated simulation-based fuzzy logic model for real-time traffic signal control. Transportmetrica, 7(4), 247–278. https://doi.org/10.1080/18128600903427645

Hoque, M. S., & Asif Imran, M. (2007). Modification of Webster’s delay formula under non-lane based heterogeneous road traffic condition. In Journal of Civil Engineering (IEB) (Vol. 35, Number 2).

Kodama, N., Harada, T., & Miyazaki, K. (2022). Traffic Signal Control System Using Deep Reinforcement Learning With Emphasis on Reinforcing Successful Experiences. IEEE Access, 10(September), 128943–128950. https://doi.org/10.1109/ACCESS.2022.3225431

Manthirikul, S., Amshala, V. T., & Jain, U. (2023). Modeling vehicular and pedestrian delays at signalized midblock crosswalk under mixed traffic conditions. Transportation Letters, 15(1), 62–75. https://doi.org/10.1080/19427867.2021.2019487

Manthirikul, S., Jain, U., & Marisamynathan, S. (2023). Pedestrian delay models for compliant & non-compliant behaviour at signalized midblock crosswalks under mixed traffic conditions. IATSS Research, 47(3), 401–415. https://doi.org/10.1016/j.iatssr.2023.08.003

Mao, T., Mihaita, A.-S., Chen, F., & Vu, H. L. (2021). Boosted Genetic Algorithm using Machine Learning for traffic control optimization. http://arxiv.org/abs/2103.08317

Musa, P., Wibowo, E. P., Musa, S. B., & Baihaqi, I. (2022). Pelican Crossing System for Control a Green Man Light with Predicted Age. MATRIK : Jurnal Manajemen, Teknik Informatika Dan Rekayasa Komputer, 21(2), 293–306. https://doi.org/10.30812/matrik.v21i2.1508

Mussone, L., & Hassan, O. el. (2025). An analysis of pedestrian crossings through deep learning models and crash data. Transportation Research Interdisciplinary Perspectives, 31(April), 101449. https://doi.org/10.1016/j.trip.2025.101449

Song, X., Yang, Q., Qu, Z., Tao, P., & Li, Z. (2015). New model for total crossing time of pedestrian platoon at a signalized crosswalk. Transportation Research Record, 2519, 67–74. https://doi.org/10.3141/2519-08

Yulianto, B., & Setiono. (2017). Analysis of Demand Responsive Pelican Crossing in Mixed Traffic Conditions. Procedia Engineering, 171, 1472–1477. https://doi.org/10.1016/j.proeng.2017.01.473

Zhang, Z., Wang, D., Liu, T., & Liu, Y. (2016). Waiting endurance time of pedestrians crossing at signalized intersections in Beijing. Transportation Research Record, 2581, 95–103. https://doi.org/10.3141/2581-12

Download

Share

COinS