A human driver model for autonomous lane changing in highways: Predictive fuzzy Markov game driving strategy
Başlık çevirisi mevcut değil.
- Tez No: 598697
- Danışmanlar: PROF. REZA LANGARI
- Tez Türü: Doktora
- Konular: Makine Mühendisliği, Mechanical Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2018
- Dil: İngilizce
- Üniversite: Texas A&M University
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 176
Özet
Özet yok.
Özet (Çeviri)
This study presents an integrated hybrid solution to mandatory lane changing problem to deal with accident avoidance by choosing a safe gap in highway driving. To manage this, a comprehensive treatment to a lane change active safety design is proposed from dynamics, control, and decision making aspects. My effort first goes on driver behaviors and relating human reasoning of threat in driving for modeling a decision making strategy. It consists of two main parts; threat assessment in traffic participants, (TV s) states, and decision making. The first part utilizes an complementary threat assessment of TV s, relative to the subject vehicle, SV , by evaluating the traffic quantities. Then I propose a decision strategy, which is based on Markov decision processes (MDPs) that abstract the traffic environment with a set of actions, transition probabilities, and corresponding utility rewards. Further, the interactions of the TV s are employed to set up a real traffic condition by using game theoretic approach. The question to be addressed here is that how an autonomous vehicle optimally interacts with the surrounding vehicles for a gap selection so that more effective performance of the overall traffic flow can be captured. Finding a safe gap is performed via maximizing an objective function among several candidates. A future prediction engine thus is embedded in the design, which simulates and seeks for a solution such that the objective function is maximized at each time step over a horizon. The combined system therefore forms a predictive fuzzy Markov game (FMG) since it is to perform a predictive interactive driving strategy to avoid accidents for a given traffic environment. I show the effect of interactions in decision making process by proposing both cooperative and non-cooperative Markov game strategies for enhanced traffic safety and mobility. This level is called the higher level controller. I further focus on generating a driver controller to complement the automated car's safe driving. To compute this, model predictive controller (MPC) is utilized. The success of the combined decision process and trajectory generation is evaluated with a set of different traffic scenarios in dSPACE virtual driving environment. Next, I consider designing an active front steering (AFS) and direct yaw moment control (DYC) as the lower level controller that performs a lane change task with enhanced handling performance in the presence of varying front and rear cornering stiffnesses. I propose a new control scheme that integrates active front steering and the direct yaw moment control to enhance the vehicle handling and stability. I obtain the nonlinear tire forces with Pacejka model, and convert the nonlinear tire stiffnesses to parameter space to design a linear parameter varying controller (LPV) for combined AFS and DYC to perform a commanded lane change task. Further, the nonlinear vehicle lateral dynamics is modeled with Takagi-Sugeno (T-S) framework. A state-feedback fuzzy H1 controller is designed for both stability and tracking reference. Simulation study confirms that the performance of the proposed methods is quite satisfactory.
Benzer Tezler
- Applications of deep reinforcement learning for advanced driving assistance systems
İleri sürüş destek sistemleri için derin pekiştirmeli öğrenme uygulamaları
MUHARREM UĞUR YAVAŞ
Doktora
İngilizce
2023
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiMekatronik Ana Bilim Dalı
DOÇ. DR. TUFAN KUMBASAR
- Microscopic traffic simulations for exploring theinteraction dynamics of connected & autonomousvehicles in merging scenarios
Birleşme senaryolarında bağlantılı ve otonomaraçların dinamik etkileşimlerini keşfetmekiçin mikroskopic simülasyon
MEHDI MALEKI
Yüksek Lisans
İngilizce
2020
Uçak Mühendisliğiİstanbul Teknik ÜniversitesiUçak ve Uzay Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ NAZIM KEMAL ÜRE
- The performance analysis of variable time gap adaptive cruise control for different algorithms with model based feedforward control structure
Model tabanlı ileri besleme kontrol yapısıyla, farklı uyarlanabilir hız sabitleyiciler için değişken zaman açıklık algoritmasının performans analizi
ONUR EVİRGEN
Yüksek Lisans
İngilizce
2021
Mekatronik Mühendisliğiİstanbul Teknik ÜniversitesiMekatronik Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ İLKER ÜSTOĞLU
- Reducing in-vehicle communication overload and enhancing efficiency in autonomous and electrical vehicles
Otonom ve elektrikli araçlarda araç içi iletişim yükünü azaltma ve etkinliğini artırma
YUNUS KAĞAN ÖZDEMİR
Yüksek Lisans
İngilizce
2024
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiElektrik Mühendisliği Ana Bilim Dalı
PROF. DR. AHMET CANSIZ
- Automated lane change decision making for autonomous vehicles using machine learning techniques
Makine öğrenmesi teknikleri ile otonom araçlarda şerit değişimine karar verme
MEHDI NASIRI
Yüksek Lisans
İngilizce
2020
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiKontrol ve Otomasyon Mühendisliği Ana Bilim Dalı
DOÇ. GÜLAY ÖKE GÜNEL