Multiagent reinforcement learning for autonomous driving in traffic zones with unsignalized intersections

Spatharis, Christos; Blekas, Konstantinos

doi:10.1080/15472450.2022.2109416

Cited by 12 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…TORCS and SUMO are demonstrated to be prominent tools for developing and testing short-term trajectory planning approaches [50], solutions that handle complex state and action spaces in a continuous domain [51], multi-agent deep reinforcement learning methods for self-driving vehicles capable of navigating through traffic networks with uncontrolled intersections [52], and many other advances or methodologies.…”

Section: State Of the Art Reviewmentioning

confidence: 99%

Deep Reinforcement Learning for Autonomous Driving in Amazon Web Services DeepRacer

Petryshyn,

Postupaiev,

Ben Bari

et al. 2024

Information

View full text Add to dashboard Cite

The development of autonomous driving models through reinforcement learning has gained significant traction. However, developing obstacle avoidance systems remains a challenge. Specifically, optimising path completion times while navigating obstacles is an underexplored research area. Amazon Web Services (AWS) DeepRacer emerges as a powerful infrastructure for engineering and analysing autonomous models, providing a robust foundation for addressing these complexities. This research investigates the feasibility of training end-to-end self-driving models focused on obstacle avoidance using reinforcement learning on the AWS DeepRacer autonomous race car platform. A comprehensive literature review of autonomous driving methodologies and machine learning model architectures is conducted, with a particular focus on object avoidance, followed by hands-on experimentation and the analysis of training data. Furthermore, the impact of sensor choice, reward function, action spaces, and training time on the autonomous obstacle avoidance task are compared. The results of the best configuration experiment demonstrate a significant improvement in obstacle avoidance performance compared to the baseline configuration, with a 95.8% decrease in collision rate, while taking about 79% less time to complete the trial circuit.

show abstract

Section: State Of the Art Reviewmentioning

confidence: 99%

Deep Reinforcement Learning for Autonomous Driving in Amazon Web Services DeepRacer

Petryshyn,

Postupaiev,

Ben Bari

et al. 2024

Information

View full text Add to dashboard Cite

show abstract

“…Several research have introduced optimization strategies to further improve intersection performance, including bilevel programming [5,6], tree search [7], and machine learning [8][9][10][11], reservation optimization [12], other strategies [13,14], yielding satisfactory schedule results. However, Lioris et al [15] found that increasing intersection capacity is feasible when vehicles traverse intersections as platoons rather than individually.…”

Section: Introductionmentioning

confidence: 99%

Game Scheduling for Vehicle Platoons at Autonomous Intersection Accommodating Priority Passage Requests

Hu,

Peng

2024

IEEE Access

View full text Add to dashboard Cite

Addressing the human-centric needs in vehicle scheduling at intersections poses a significant challenge. In this paper, we propose a game scheduling method to address the problem of platoon scheduling while considering priority passage requests. Firstly, we construct a coalition game-based platoon scheduling model. Within this model, a kernel function searches for the optimal coalition group within the kernel set to ensure that participants, such as platoons, gain higher payoffs by joining the coalition than they would individually. The coalition payoff function determines the optimal passage order. Secondly, we design a baseline strategy where the payoff and action obtained by the priority passage platoon in the coalition game serve as the baseline values. Finally, based on this strategy, we develop a bargaining game model where the priority passage platoon acts as the bargainer, while other platoons act as the other party, with baseline payoff and action serving as constraints. The Nash equilibrium identifies the optimal passage order for the priority passage platoon. We conduct 3 experiments: experiment 1 evaluates the effectiveness of the coalition game model; experiment 2 tests the effectiveness of the bargaining game model; and experiment 3 assesses the effectiveness of the game scheduling method in reducing vehicle travel costs and compares it with existing methods. The results demonstrate that our proposed game scheduling method not only reduces the travel costs of the priority passage platoon but also minimizes its impact on other vehicles to the greatest extent possible.INDEX TERMS Autonomous intersection, baseline payoff, game scheduling, priority passage request, vehicle platoon.

show abstract

“…In MARL, the interaction is usually modeled as a Stochastic Game(Markov Game) [100,101] instead of MDP, so collective policy can be learnt for all the agents. For example in [102][103][104] MARL is used for multiple AVs planning and decision-making. In [105,106] MARL is utilised to train altruistic autonomous agents to drive safely but also master coordination among agents.…”

Section: Game Theoretic -Game Vs Mdpmentioning

confidence: 99%

Interactive decision-making for autonomous vehicles based on a theory of order

Cai

View full text Add to dashboard Cite

Autonomous driving is challenging, because understanding how human-being makes decisions can be difficult. To build such a machine that substitutes for human to make decisions, one essential is the capability of interacting with other road users.Currently autonomous vehicles lack such an understanding, thus over conservative behaviors must be taken, which leads to compromised performances. The subject of this thesis is the development of the Theory of Order and the order-based approaches as the tool, to understand human behavior, hence enable autonomous vehicles to make interactive decisions. The theory is a mathematical framework for describing the strategies a human-like mind would take during the traffic interactions modelled as a game. It aims to cut the issue of multi-equilibria and incomplete information that makes the game difficult to solve. The meaning of Order is threefold: it describes a set of steady states that generally exists in the interactions, as a subjection relation of behaviors among the players, which is not precisely known but could be dynamically approximated through a process of order establishment; it also indicates the sequence in which the traffic space is occupied by players, and how objective and random differences would lead to the inevitability of the final subjection relation; and it's a strategy profile regarding player's aggressiveness that establish a high level equilibrium in the sense of considering driving as infinite repeated game, therefore provides a solution to multi-equilibria.The order establishment could be achieved by players adjusting their intentions constantly according to the order-strategy, which is explored both on a theoretical and on an algorithmic level. By conducting various simulations in both structured space (unsignalized intersection) and unstructured space (Arc de Triomphe roundabout), the approach is tested and verified in completing specific tasks, including unprotected left-turn, merging, or free interaction in proximity. Compared to the baseline methods, order-based approach shows its advantages in several aspects: promoting both individual and system efficiency, generating collision free maneuver, being capable of taking initiative actions in a safe manner. Exhaustive test reveals that system level performance would be promoted if players are relatively xiii xiv equivalent in aggressiveness, which indicates that AV should be designed adaptive when dealing with different groups of people. In general, this work propose an innovative framework to interpret and analyse traffic interactions and developed algorithm that improve performances in safety and efficiency for the autonomous vehicles to interact with humans.

show abstract

Multiagent reinforcement learning for autonomous driving in traffic zones with unsignalized intersections

Cited by 12 publications

References 23 publications

Deep Reinforcement Learning for Autonomous Driving in Amazon Web Services DeepRacer

Deep Reinforcement Learning for Autonomous Driving in Amazon Web Services DeepRacer

Game Scheduling for Vehicle Platoons at Autonomous Intersection Accommodating Priority Passage Requests

Interactive decision-making for autonomous vehicles based on a theory of order

Contact Info

Product

Resources

About