Autonomous Driving Strategies at Intersections: Scenarios, State-of-the-Art, and Future Outlooks

Abstract

Due to the complex and dynamic character of intersection scenarios, the autonomous driving strategy at intersections has been a difficult problem and a hot point in the research of intelligent transportation systems in recent years. This paper gives a brief summary of state-of-the-art autonomous driving strategies at intersections. Firstly, we enumerate and analyze common types of intersection scenarios, corresponding simulation platforms, as well as related datasets. Secondly, by reviewing previous studies, we have summarized characteristics of existing autonomous driving strategies and classified them into several categories. Finally, we point out problems of the existing autonomous driving strategies and put forward several valuable research outlooks.

Urban Intersection Simulation and Verification via Deep Reinforcement Learning Algorithms

“…A simulation environment for training the RL algorithms had the challenge and limitations of simulating an actual situation. Wei et al [4] discussed that most automatic driving strategies had not included the more complex pedestrians to simulate more dynamics. It can be challenging to train the algorithms to have more safety in this situation.…”

“…The traditional method of a traffic light might rely on fixed time cycles to be challenging to solve the problem of dynamic traffic environment in the different periods of intersection [3]. The different types of intersections had a complex environment and connected the other main road in the urban area, including the crossroad, roundabout, and deformed intersections [4]. As the main urban transport road and a more complex situation, the intersection has undertaken more pressure on the designated traffic time.…”

“…Furthermore, the complex environment and conflict points could cause a greater probability between the different collision vehicles in the collision [4]. Due to solve the urban intersection problem, automatic driving has become a method based on Computer Vision and Reinforcement Learning to build the digital urban control system.…”

Autonomous Driving Decision Algorithm for Complex Multi-Vehicle Interactions: An Efficient Approach Based on Global Sorting and Local Gaming

“…There are two main methodologies to solve the intersection interaction problem, i.e. multi-vehicle collaboration (MVC) and single-vehicle intelligence (SVI) [3].…”

“…On the other hand, to find SVI-based strategies it is usually necessary to obtain the traffic environment information via on-board sensors, and then to estimate or predict the other participants' driving style and future intention/trajectory [3]. The existing literature on SVI-based unsignalized intersection decision mainly include rule-based [9], decision tree based, learning-based and game-theoretic approaches.…”