Autonomous driving has witnessed rapid advancement;however,ensuring safe and efficient driving in intricate scenarios remains a critical challenge.In particular,traffic roundabouts bring a set of challenges to autonom...Autonomous driving has witnessed rapid advancement;however,ensuring safe and efficient driving in intricate scenarios remains a critical challenge.In particular,traffic roundabouts bring a set of challenges to autonomous driving due to the unpredictable entry and exit of vehicles,susceptibility to traffic flow bottlenecks,and imperfect data in perceiving environmental information,rendering them a vital issue in the practical application of autonomous driving.To address the traffic challenges,this work focused on complex roundabouts with multi-lane and proposed a Perception EnhancedDeepDeterministic Policy Gradient(PE-DDPG)for AutonomousDriving in the Roundabouts.Specifically,themodel incorporates an enhanced variational autoencoder featuring an integrated spatial attention mechanism alongside the Deep Deterministic Policy Gradient framework,enhancing the vehicle’s capability to comprehend complex roundabout environments and make decisions.Furthermore,the PE-DDPG model combines a dynamic path optimization strategy for roundabout scenarios,effectively mitigating traffic bottlenecks and augmenting throughput efficiency.Extensive experiments were conducted with the collaborative simulation platform of CARLA and SUMO,and the experimental results show that the proposed PE-DDPG outperforms the baseline methods in terms of the convergence capacity of the training process,the smoothness of driving and the traffic efficiency with diverse traffic flow patterns and penetration rates of autonomous vehicles(AVs).Generally,the proposed PE-DDPGmodel could be employed for autonomous driving in complex scenarios with imperfect data.展开更多
Existing studies on modern roundabouts performance are mostly based on data fron： singe lane roundabouts that are not heavily congested. For planners and designers interested in building multilane roundabouts for int...Existing studies on modern roundabouts performance are mostly based on data fron： singe lane roundabouts that are not heavily congested. For planners and designers interested in building multilane roundabouts for intersections with potential growth i~ future traffic, there has been a lack of existing studies with field data that provide reference values in terms of capacity and delay measurements. With the intent of providing such reference values, a case study was conducted by using the East DowlinC Road Roundabouts in Anchorage, Alaska, which are currently operating with extensive queues during the evening peak hours. This research used multiple video camcorders t（ capture vehicle turning movements at the roundabouts as well as the progressior~ of vehicle queues at the roundabout entrance approaches. With these video records, the number of vehicles in the queues can be accurately counted in any single minute during the peak hours. This study shows that unbalanced entrance flow patterns （i.e., ~ne entrance has significant higher flow than others） can intensify the queue and delay fo., the overall roundabouts. Then various software packages including RODEL, SIDRA and VISSIM were used to estimate several performance measurements, such as capacity. queue length, and delay, compared with the collected field data. With the comparison, it is found that all the three software packages overestimate multi-lane roundabout ca pacity before calibration. With default parameters, SIDRA and VISSIM tend to underes timate delays and queue lengths for the multi-lane roundabouts under congestion, while RODEL results in higher delay and queue length estimations at most of the entrance approaches.展开更多
基金supported in part by the projects of the National Natural Science Foundation of China(62376059,41971340)Fujian Provincial Department of Science and Technology(2023XQ008,2023I0024,2021Y4019),Fujian Provincial Department of Finance(GY-Z230007,GYZ23012)Fujian Key Laboratory of Automotive Electronics and Electric Drive(KF-19-22001).
文摘Autonomous driving has witnessed rapid advancement;however,ensuring safe and efficient driving in intricate scenarios remains a critical challenge.In particular,traffic roundabouts bring a set of challenges to autonomous driving due to the unpredictable entry and exit of vehicles,susceptibility to traffic flow bottlenecks,and imperfect data in perceiving environmental information,rendering them a vital issue in the practical application of autonomous driving.To address the traffic challenges,this work focused on complex roundabouts with multi-lane and proposed a Perception EnhancedDeepDeterministic Policy Gradient(PE-DDPG)for AutonomousDriving in the Roundabouts.Specifically,themodel incorporates an enhanced variational autoencoder featuring an integrated spatial attention mechanism alongside the Deep Deterministic Policy Gradient framework,enhancing the vehicle’s capability to comprehend complex roundabout environments and make decisions.Furthermore,the PE-DDPG model combines a dynamic path optimization strategy for roundabout scenarios,effectively mitigating traffic bottlenecks and augmenting throughput efficiency.Extensive experiments were conducted with the collaborative simulation platform of CARLA and SUMO,and the experimental results show that the proposed PE-DDPG outperforms the baseline methods in terms of the convergence capacity of the training process,the smoothness of driving and the traffic efficiency with diverse traffic flow patterns and penetration rates of autonomous vehicles(AVs).Generally,the proposed PE-DDPGmodel could be employed for autonomous driving in complex scenarios with imperfect data.
基金sponsored by Alaska University Transportation Center(AUTC,No.RR08.08)Alaska Department of Transportation(AK DOT)
文摘Existing studies on modern roundabouts performance are mostly based on data fron： singe lane roundabouts that are not heavily congested. For planners and designers interested in building multilane roundabouts for intersections with potential growth i~ future traffic, there has been a lack of existing studies with field data that provide reference values in terms of capacity and delay measurements. With the intent of providing such reference values, a case study was conducted by using the East DowlinC Road Roundabouts in Anchorage, Alaska, which are currently operating with extensive queues during the evening peak hours. This research used multiple video camcorders t（ capture vehicle turning movements at the roundabouts as well as the progressior~ of vehicle queues at the roundabout entrance approaches. With these video records, the number of vehicles in the queues can be accurately counted in any single minute during the peak hours. This study shows that unbalanced entrance flow patterns （i.e., ~ne entrance has significant higher flow than others） can intensify the queue and delay fo., the overall roundabouts. Then various software packages including RODEL, SIDRA and VISSIM were used to estimate several performance measurements, such as capacity. queue length, and delay, compared with the collected field data. With the comparison, it is found that all the three software packages overestimate multi-lane roundabout ca pacity before calibration. With default parameters, SIDRA and VISSIM tend to underes timate delays and queue lengths for the multi-lane roundabouts under congestion, while RODEL results in higher delay and queue length estimations at most of the entrance approaches.
