Araç-yol etkileşimi ortamında sinyalize kavşaklarda kişi gecikmesini enküçükleyen bir algoritma
An algorithm minimizing person delay in isolated signalized intersections in vehicle-to-infrastructure environment
- Tez No: 932141
- Danışmanlar: DOÇ. DR. MURAT ERGÜN
- Tez Türü: Doktora
- Konular: Ulaşım, Transportation
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2025
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Lisansüstü Eğitim Enstitüsü
- Ana Bilim Dalı: İnşaat Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Ulaştırma Mühendisliği Bilim Dalı
- Sayfa Sayısı: 130
Özet
Sinyalize kavşaklarda toplu taşıma önceliği, işletme maliyetlerinin düşürülmesinde ve toplu taşıma kullanımının teşvik edilmesinde önemli bir rol oynamaktadır. Geleneksel olarak kavşaklardaki sinyal sistemleri yollarda veya araçlarda bulunan sensörler aracılığıyla toplu taşıma araçlarına kavşaklarda geçiş önceliği sağlarlar. Ancak bu sadece toplu taşıma araçlarıyla sınırlıdır. Toplu taşıma araçları ve özellikle otobüsler şehirlerde insan hareketliliğini artırmakta en önemli unsurlar olsa da özel araçlar ve ara toplu taşıma modları da şehirlerin bir gerçeği olup onları da hareketliliği artırma konusunda verimli kullanmaya teşvik etmek gerekmektedir. Bazı ülkelerde uygulanan otoyollarda yüksek doluluklu araçlara özel şerit tahsis edilmesi uygulamasında olduğu gibi kentsel alanlarda bulunan kavşaklarda sadece resmi toplu taşıma araçlarına değil, ikiden fazla yolcusu olan her türlü araca sinyalizasyonda öncelik verilebilir. Bugüne kadar algılama teknolojilerinin kontrol yöntemlerinin ihtiyaç duyduğu verileri sağlayamaması dolayısıyla hayata geçirilemeyen böylesi bir uygulama araç-yol etkileşimi gibi gelişen yeni teknolojiler ile etkin bir şekilde hayata geçirilebilir. Ayrıca araç-yol etkileşimi teknolojisi, yolcu sayma teknolojileri ile birleştirilerek bir kavşakta yolcu başına düşen gecikmeler hesaplanabilmektedir. Tez çalışmasının amacı izole ışıklı kavşaklarda taşıtlara içlerindeki yolcu sayısı ile orantılı bir şekilde öncelik sağlayan ve bu şekilde kişi başına gecikmeyi enküçükleyen bir algoritma sunmaktır. Bir başka ifadeyle, sunulan algoirtma ile geleneksel toplu taşıma önceliği uygulamalarının sadece otoüslere sağladığı avantajlardan ara toplu taşıma araçları ve yüksek doluluklu özel otomobillerin de faydalanmasına imkan verilmesi hedeflenmiştir. Bu amaçla, bu çalışma, sinyal denetleyicisi ile iletişim için yolcu sayımı teknolojilerini ve araç-yol etkileşimini kullanır. Bu çalışma, bağlantılı araç teknolojisinden yararlanarak izole kavşaklarda insan hareketliliğini artırmak için bir algoritma sunmaktadır. Önerilen algoritma, araçlardaki yolcu sayısını dikkate alarak parçacık sürü optimizasyonu yöntemini kullanarak optimum sinyal sürelerini üretmektedir. İstanbul Güngören'deki bir kavşağın gerçek sayım ve sınıflarma verileri kullanılarak benzetim ortamında farklı zaman dilimleri için gerçekleştirilen testler, önerilen algoritmanın geleneksel toplu taşıma önceliği algoritmalarının başarımını çok aştığını göstermektedir. Sınırlı işlem gücüne sahip gömülü bir sistemde uygulanmaya uygun yüksek başarımlı olarak tasarlanan algoritma, eniyileştirilmiş bir sabit süreli çoklu planlı kontrol algoritmasına kıyasla yolcubaşına gecikmede %14'lük bir iyileşme ve geleneksel bir otobüs öncelikli kontrol algoritmasına kıyasla yolcu başına gecikmede %8'lik bir iyileşme sağlar.
Özet (Çeviri)
Vehicle ownership is an important factor for traffic congestion in urban areas. The sole solution for the traffic congestion is to promote public transport (PT) usage. Private vehicles have very low passenger-carrying capacity and utilization compared to public transport. For example, the average number of passengers per private vehicle in Istanbul was determined to be 1.57 in the Istanbul Transportation Master Plan, while the average number of passengers per vehicle for public transportation (bus) was determined as 30.4. The aim of a traffic control system should be to increase the mobility of people rather than vehicles. High passenger-carrying capacity makes buses ideal vehicles for a mobility-friendly city. As a result, bus and tram priority has been applied in traffic control systems with more or less sophistication in almost every major city in Europe. However, signal priority applications are restricted only to public transport vehicles. The idea of High Occupancy Vehicles (HOV) lane applications in interurban roads can be converted to urban areas and priority can be given not only to official transit vehicles but also to any vehicle with more than two passengers. In this regard, emerging technologies like vehicle-to-infrastructure communication (V2I) may assist these studies. In addition, V2I technology can be combined with passenger count technologies, thus delays per passengers at an intersection can be calculated. The aim of the thesis study is to propose an algorithm that prioritizes vehicles at signalized intersections in proportion to the number of passengers they carry, thereby minimizing per capita delay. In other words, the proposed algorithm seeks to extend the benefits of traditional public transportation priority systems, which typically favor buses, to intermediate public transportation vehicles and high-occupancy private cars as well. For this purpose, this study uses passenger count technologies and V2I for communicating with the signal controller. This study introduces an algorithm for increasing human mobility in isolated intersections by taking advantage of connected-vehicles technology. The proposed algorithm named TS-PDM generates the best signal times by taking into account the number of passengers in vehicles. The proposed algorithm runs on the intersection controller unit with Roadside Unit (RSU). Vehicles' On-board Units (OBU) periodically report location, temporary ID, speed, and number of passengers in the vehicle with a time stamp. In order for the algorithm to work effectively, the system needs to know vehicle information in sections, starting from the stop line and ending at the various distances. Traffic demand at the intersection is determined by observing those sections from every direction. The optimum phase green times that minimize delays at the intersection are determined by evaluating the traffic volumes from all approaches using a heuristic algorithm such as particle swarm optimization (PSO), genetic algorithms (GA), ant colony optimization (ACO), etc. The delay-optimization problem is not an NP-hard problem. Applying it in an embedded system with limited processing power or in a central system consisting of many nodes might require high performance and algorithmic simplicity. This study uses the PSO algorithm. PSO is a concept first discussed by Eberhart and Kennedy and was inspired from the social behaviors of bird flocks and fish schools for optimizing nonlinear functions. Implementing the PSO algorithm is simple and has become increasingly common in many different areas including the transportation sector. PSO and GA have been compared within the scope of this study, and PSO is found to converge to optimal values much quicker than GA. The PSO algorithm converges on a better solution (with an error 5.88% the size of GA) using the same number of particles (population size) and fewer iterations (up to 28.6% less than GA). The total passenger-delay values for the directions are used as a fitness function in the PSO algorithm. To obtain the total delay value for each direction, the vehicle delay is obtained from the HCM-2000 delay formula and multiplied by the total number of passengers. In order to keep up with the variability of traffic demand, the TS-PDM algorithm includes some predetermined parameters as well as instantly changing parameters. The minimum phase length is fixed and pre-determined by the user. The algorithm does not run phases below the minimum time. Meanwhile, phase-length maximum limits are very important for adapting to changes in traffic demand. A more sophisticated approach is needed in order to minimize per-person delays. The traffic volumes measured in the previous cycle help estimate traffic volume in the next cycle but in reality, do not guarantee what the volume will actually be. Therefore, the control system should be able to generate a response when the traffic demand does not match the estimate. The TS-PDM system terminates the relevant green early when the traffic demand from one direction is not as expected, thus not unnecessarily delaying the vehicles in the crossing directions. The TS-PDM control method takes vehicle and passenger counts at 120-sec intervals using point detectors about 300 m from the stop line. At the beginning of each phase, the number of vehicles and passengers in the last count is converted into hourly volumes and input into the PSO algorithm; the green time of the phase to be started is calculated and applied. For testing the proposed algorithm, a two-phase intersection has been selected at Gungoren in Istanbul. In order to see the effects of developments in the algorithm more clearly, a basic-structured intersection with very little pedestrian movement has been chosen, with a PTV-Vissim file being created to simulate the selected and nearby intersections as the traffic network. Traffic volumes for seven different vehicle classes (automobile, pickup, taxi, minibus/van, paratransit vehicle, bus, and truck) for three time zones (morning peak hours, off-peak hours, evening peak hours) have been collected from field at 15-minute intervals and entered into the simulaiton file. Passengers were assigned with a Poisson distribution function for each vehicle produced in the simulation based on the average number of passengers for each vehicle class from the Istanbul Transportation Master Plan. The number of passengers assigned to each object vehicle in the simulation has then been used to calculate delays later. Since, this study focuses on the signal control algorithm rather than the communication technology itself a roug simulation model of vehicle-to-vehicle communication is. The software model for a traffic simulation program has been developed in the program PTV-Vissim for this purpose. The simulation determines the likelihood that any vehicle around a coordinate set in this model will receive a wireless signal. For comparison purposes, two signal controls have been conducted apart from the proposed signal control. One of these methods is one that represents the time-dependent multi-plan structure and called TDMC. In this method, phase times are calculated offline beforehand for different traffic volumes. These fixed phase times are then activated according to the time of the day. The second of these methods is designed to represent traditional bus priority signal control systems and called BPSC. The bus-priority control algorithm applies the options of“early green”and "green time extension. Two-hour simulations were run 10 times in order to find out whether the proposed TS-PDM method had superiority over other conventional methods; average delay values, emission, and stop counts have been recorded. The expected result is a significant reduction in per-person delay values compared to the TDMC and BPSC methods. Delay measurements have been obtained using a custom script and by measuring travel times in 200m-long road segments; the number of passengers has been taken into consideration because the per-person delay (including passengers in private vehicles) is not measurable using standard delay measurements in PTV-Vissim. The intent of this paper has been to propose an algorithm for signalized intersection in V2I system to minimize person delays. To achieve this goal, particle swarm optimization algorithm is adapted for designing the proposed control algorithm of the traffic signal system of an arterial road. The effectiveness of the proposed control algorithm was then compared for Time-Dependent Multiplan Control and Bus Priority Signal Control in undersaturated conditions. For the numerical sample, the VISSIM (representing real traffic conditions) simulations were used in the two-phase signal system. The results from the performance of the numerical case study simulation are stated below: Delays per passenger are decreased by 14.09% compared to TDMC and by 8.33% compared to BPSC. - Delays per vehicle are reduced by 6.20% compared to TDMC and by 10.78% compared to BPSC by vehicle. - Delays per HOV vehicles (vehicles with five or more passengers) are reduced by 8.24% compared to TDMC and by 14.54% compared to BPSC. - Number of stops and fuel consumption perform better than the other methods when the intersection signal is controlled by TS-PDM. - When increasing the proportion of buses, the BPSC method controls bus delays better than TS-PDM. However, while the BPSC method increases delays per person, delays per bus with TS-PDM are reduced by 27.80% compared to TDMC. However, TS-PDM has an increase of 6.18% compared to BPSC. As can be understood from the results, a significant improvement occurs in per-person delays without causing an increase in vehicle delays using the TS-PDM method. Moreover, unlike the bus priority systems, the system increases the mobility of people, not buses in the city. This study was conducted at an isolated two-phase intersection and will encourage human-focused innovative approaches to signal optimization in the coming years. Furthermore, the method can be generalized to junctions and arterials with a large number of phases. Better delay scores can be achieved by modifying the method for application in arterial and by coordinating.
Benzer Tezler
- Exploring hand interfaces within new digital musical instruments research
El arayüzlerinin yeni sayısal müzik çalgıları kapsamında incelenmesi
SAİR SİNAN KESTELLİ
Doktora
İngilizce
2019
Eğitim ve Öğretimİstanbul Teknik ÜniversitesiMüzik Ana Bilim Dalı
PROF. DR. CAN KARADOĞAN
DR. REUBEN DELAUTOUR
- Uyarlamalı araç takip sistemlerinde model öngörülü kontrol yöntemleri: Karşılaştırmalı bir çalışma
Model predictive control approaches for adaptive cruise control systems: A comparative study
UMUT KARAPINAR
Yüksek Lisans
Türkçe
2017
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiKontrol ve Otomasyon Mühendisliği Ana Bilim Dalı
PROF. DR. AFİFE LEYLA GÖREN
- Smart waste management techniques and their performance under various scenarios
Akıllı atık yönetim teknikleri ve farklı senaryolardaki başarımları
DAMLA GÜLEN
Yüksek Lisans
İngilizce
2019
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
PROF. DR. SEMA FATMA OKTUĞ
- 1H-MRSI of the deep gray matter structures in patients with amyotrophic lateral sclerosis
Amyotrofik lateral skleroz hastalarında derin gri madde yapılarının multi voksel proton manyetik rezonans spektroskopisi
MERYEM TORLAK
Yüksek Lisans
İngilizce
2024
Biyomühendislikİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
PROF. DR. İSA YILDIRIM
PROF. DR. ESİN ÖZTÜRK IŞIK
- Enhancement of situational awareness in physical security using mixed reality
Fiziksel güvenlikte karma gerçeklikle durumsal farkındalığın artırılması
NAZIM YİĞİT KAVASOĞLU
Yüksek Lisans
İngilizce
2023
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiBilgisayar Mühendisliği Ana Bilim Dalı
DOÇ. DR. GÖKHAN İNCE