Intentional non-yielding intelligent driver model development for cut-in & cut-out situations considering interaction with surrounding vehiclesCut-in 및 Cut-out 상황에서의 주변 차량과의 상호 작용을 고려한 의도적 비양보 지능형 운전자 모델 개발

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 526
  • Download : 0
Planning or decision making algorithms of autonomous driving technology are often verified in a simulation environment before being tested in a real road environment. With the development of simulators, traffic flow from a spare road to a densely packed road is being implemented well. However, driver models that represent human behavior while interacting with other vehicles are still lacking. The decision is mostly made in the direction of yielding to cut in vehicles, and there is a difference in degree. For this reason, trade off between the efficiency and safety are being shown in poor level which leads to the poor performance of validating or training the algorithm of motion planning. Some models that represent errors of the human driver have been developed but do not represent the drivers that act offensively with their intention. This paper focuses on the driver model for highways and simulation environments. Single or multiple lane changes occur on highways and these kinds of lane changing algorithms are being tested in the simulation environment. However, with these kinds of driver models, the trade off between the lane change success rate and collision avoidance rate is weak and often shows high performance regardless of t he complexity of the traffic. Considering many of the motion planning researches are done with a learning based method, the training environment does not fully represent the real road environment, thus ending up in a failure in a real road test. The propos ed driver model can both choose to yield and non yield depending on the parameters we set, and the degree of aggressiveness can also be set. This model can be used as an intentionally non yielding aggressive driver, as well as a common car following model in cut in simulation. Helping to test many algorithms in harsh traffic conditions and helping learning based models to be trained for handling unfriendly drivers on road.
Advisors
Kum, Dong Sukresearcher금동석researcher
Description
한국과학기술원 :미래자동차학제전공,
Publisher
한국과학기술원
Issue Date
2023
Identifier
325007
Language
eng
Description

학위논문(석사) - 한국과학기술원 : 미래자동차학제전공, 2023.2,[iv, 44 p. :]

Keywords

Autonomous driving▼aDriver model▼aSimulation▼aLane change▼aTraffic flow▼aTrade-off; 자율주행▼a운전자 모델▼a시뮬레이션▼a차선 변경▼a교통 흐름▼a상충관계

URI
http://hdl.handle.net/10203/308331
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1032367&flag=dissertation
Appears in Collection
PD-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0