scholarly journals Connected autonomous vehicles for improving mixed traffic efficiency in unsignalized intersections with deep reinforcement learning

2021 ◽  
Vol 1 ◽  
pp. 100017
Author(s):  
Bile Peng ◽  
Musa Furkan Keskin ◽  
Balázs Kulcsár ◽  
Henk Wymeersch
Keyword(s):  
Reinforcement Learning ◽  
Autonomous Vehicles ◽  
Mixed Traffic ◽  
Traffic Efficiency ◽  
Unsignalized Intersections

scholarly journals Influence of Lane Policies on Freeway Traffic Mixed with Manual and Connected and Autonomous Vehicles

2020 ◽  
Vol 2020 ◽  
pp. 1-20
Author(s):  
Xuedong Hua ◽  
Weijie Yu ◽  
Wei Wang ◽  
Wenjie Xie
Keyword(s):  
Autonomous Vehicles ◽  
Traffic Management ◽  
Heterogeneous Traffic ◽  
Mixed Traffic ◽  
Freeway Traffic ◽  
Average Speed ◽  
Traffic Efficiency ◽  
Management Measures ◽  
Traffic Volumes ◽  
Ca Model

Connected and autonomous vehicles (CAVs) have become the highlights of traffic. Researchers in this field have proposed various traffic management measures to enhance the capacity and efficiency of traffic with CAVs, especially mixed traffic of CAVs and manual vehicles (MVs). Exclusive lane setting is included. However, exclusive lane policy-related researches for mixed traffic of CAVs and MVs were very limited, and the influence of number and location of exclusive lanes on the mixed traffic was unclear. To fill this gap, this paper aims to study the influence of different exclusive lane policies on mixed traffic and provide recommended lane policies under various traffic volumes and CAV penetration rates. Freeways with two lanes and three lanes in a single direction were taken into consideration, and sixteen lane policies were proposed. Then different lane policies were simulated with a new proposed cellular automata (CA) model, and properties including flux, average speed, and CAVs degradation were analyzed to evaluate the traffic efficiency of each lane policy. The results show that CAV exclusive lanes can improve the capacity, while MV exclusive lanes seem helpless for capacity improvement. Seven lane policies, including GC, GM, and CM for two-lane freeways and GCG, CGC, and CCM for three-lane freeways, outperform the others in terms of average speed. In addition, exclusive lanes can reduce the probability that CAVs degenerate to AVs. Our findings may help to optimize freeways’ lane policies and improve the efficiency of heterogeneous traffic mixed with CAVs and MVs.

Pricing of Connected and Autonomous Vehicles in Mixed-Traffic Networks

2021 ◽  
pp. 036119812098585
Author(s):  
Paolo Delle Site
Keyword(s):  
Autonomous Vehicles ◽  
Marginal Cost ◽  
User Behavior ◽  
Social Planner ◽  
Traffic Networks ◽  
Mixed Traffic ◽  
The Third ◽  
First Case ◽  
Marginal Cost Pricing ◽  
The Social

For networks with human-driven vehicles (HDVs) only, pricing with arc-specific tolls has been proposed to achieve minimization of travel times in a decentralized way. However, the policy is hardly feasible from a technical viewpoint without connectivity. Therefore, for networks with mixed traffic of HDVs and connected and autonomous vehicles (CAVs), this paper considers pricing in a scenario where only CAVs are charged. In contrast to HDVs, CAVs can be managed as individual vehicles or as a fleet. In the latter case, CAVs can be routed to minimize the travel time of the fleet of CAVs or that of the entire fleet of HDVs and CAVs. We have a selfish user behavior in the first case, a private monopolist behavior in the second, a social planner behavior in the third. Pricing achieves in a decentralized way the social planner optimum. Tolls are not unique and can take both positive and negative values. Marginal cost pricing is one solution. The valid toll set is provided, and tolls are then computed according to two schemes: one with positive tolls only and minimum toll expenditure, and one with both tolls and subsidies and zero net expenditure. Convergent algorithms are used for the mixed-behavior equilibrium (simplicial decomposition algorithm) and toll determination (cutting plane algorithm). The computational experience with three networks: a two-arc network representative of the classic town bypass case, the Nguyen-Dupuis network, and the Anaheim network, provides useful policy insight.

scholarly journals An Efficiency Enhancing Methodology for Multiple Autonomous Vehicles in an Urban Network Adopting Deep Reinforcement Learning

2021 ◽  
Vol 11 (4) ◽  
pp. 1514 ◽  
Author(s):  
Quang-Duy Tran ◽  
Sang-Hoon Bae
Keyword(s):  
Reinforcement Learning ◽  
Traffic Congestion ◽  
Autonomous Vehicles ◽  
Penetration Rate ◽  
Autonomous Vehicle ◽  
Effective Means ◽  
Urban Network ◽  
Learning Agents ◽  
Policy Optimization ◽  
The Impact

To reduce the impact of congestion, it is necessary to improve our overall understanding of the influence of the autonomous vehicle. Recently, deep reinforcement learning has become an effective means of solving complex control tasks. Accordingly, we show an advanced deep reinforcement learning that investigates how the leading autonomous vehicles affect the urban network under a mixed-traffic environment. We also suggest a set of hyperparameters for achieving better performance. Firstly, we feed a set of hyperparameters into our deep reinforcement learning agents. Secondly, we investigate the leading autonomous vehicle experiment in the urban network with different autonomous vehicle penetration rates. Thirdly, the advantage of leading autonomous vehicles is evaluated using entire manual vehicle and leading manual vehicle experiments. Finally, the proximal policy optimization with a clipped objective is compared to the proximal policy optimization with an adaptive Kullback–Leibler penalty to verify the superiority of the proposed hyperparameter. We demonstrate that full automation traffic increased the average speed 1.27 times greater compared with the entire manual vehicle experiment. Our proposed method becomes significantly more effective at a higher autonomous vehicle penetration rate. Furthermore, the leading autonomous vehicles could help to mitigate traffic congestion.

scholarly journals Capacity of Unsignalized Intersections under Mixed Traffic Conditions

2011 ◽  
Vol 16 ◽  
pp. 676-685 ◽  
Author(s):  
Joewono Prasetijo ◽  
Mehdi Hossein Pour ◽  
Seyed Mohammad Reza Ghadiri
Keyword(s):  
Mixed Traffic ◽  
Traffic Conditions ◽  
Unsignalized Intersections

scholarly journals Deep reinforcement learning based control for Autonomous Vehicles in CARLA

2022 ◽  
Author(s):  
Óscar Pérez-Gil ◽  
Rafael Barea ◽  
Elena López-Guillén ◽  
Luis M. Bergasa ◽  
Carlos Gómez-Huélamo ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Autonomous Vehicles ◽  
Autonomous Vehicle ◽  
Vehicle Control ◽  
Data Sources ◽  
Simulation Environment ◽  
Urban Simulation ◽  
Policy Gradient ◽  
Almost All ◽  
Control Layer

AbstractNowadays, Artificial Intelligence (AI) is growing by leaps and bounds in almost all fields of technology, and Autonomous Vehicles (AV) research is one more of them. This paper proposes the using of algorithms based on Deep Learning (DL) in the control layer of an autonomous vehicle. More specifically, Deep Reinforcement Learning (DRL) algorithms such as Deep Q-Network (DQN) and Deep Deterministic Policy Gradient (DDPG) are implemented in order to compare results between them. The aim of this work is to obtain a trained model, applying a DRL algorithm, able of sending control commands to the vehicle to navigate properly and efficiently following a determined route. In addition, for each of the algorithms, several agents are presented as a solution, so that each of these agents uses different data sources to achieve the vehicle control commands. For this purpose, an open-source simulator such as CARLA is used, providing to the system with the ability to perform a multitude of tests without any risk into an hyper-realistic urban simulation environment, something that is unthinkable in the real world. The results obtained show that both DQN and DDPG reach the goal, but DDPG obtains a better performance. DDPG perfoms trajectories very similar to classic controller as LQR. In both cases RMSE is lower than 0.1m following trajectories with a range 180-700m. To conclude, some conclusions and future works are commented.

Sign in / Sign up

Export Citation Format

Share Document