Decision making in unsignalized traffic intersection can be challenging for autonomous vehicles. An autonomous system only depends on predefined street priorities and considering other cars as moving objects will cause the car to freeze and fail the maneuverer. Human drivers influence the cooperation of other drivers to evade such deadlock situations and convince others to change their behavior. Decision making processes should reason about the interaction with other drivers and anticipate a wide range of driver behaviors. In this paper, we present a reinforcement learning approach (Q-LEARNING) to learn how to interact with drivers with different situation. We enhanced the performance of Q-LEARNING algorithms by maintaining a belief over the level of cooperation of other drivers. We show that our drives successfully learns how to navigate with less deadlocks than without learning.
Dilemma Game at Unsignalized Intersection with Reinforcement Learning
Lect. Notes in Networks, Syst.
International Conference On Big Data and Internet of Things ; 2021 ; Mohammed V University in Rabat, Morocco March 17, 2021 - March 18, 2021
Proceedings of the 5th International Conference on Big Data and Internet of Things ; Chapter : 36 ; 481-491
2022-07-03
11 pages
Article/Chapter (Book)
Electronic Resource
English
Unsignalized Intersection Simulation Model InterSim
British Library Conference Proceedings | 1997
|Dirty-Face Game Analysis on Mixed Traffic Flow at Unsignalized Intersection
Trans Tech Publications | 2011
|