Scalable self-adaptive traffic management in urban networks using FRVRL for optimizing traffic rules and road configurations

Engineering notes

Key topics: autonomous driving, control. See the paper for implementation details and experimental results.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

In recent decades with the increase in population, many cities find themselves in the problem of lacking sufficient facilities for building new roads to handle traffic congestion problems. One of the ways to address this problem is to optimize the existing traffic rules to distribute the traffic congestion throughout the city. For this reason, in this study, the authors proposed using an RL-based management strategy for controlling traffic through urban traffic rules. This method was implemented on a reduced graph structure of four different cities: London, Paris, Ankara, and New York, and evaluated across four traffic demand scenarios ranging from light to extremely heavy congestion. The proposed framework is further validated through a comparative analysis against an IoT-enabled adaptive traffic management framework on the London road network, demonstrating consistent improvements in average velocity, congestion flow, and arrival time reduction across all tested conditions. The results show good traffic management capability in reducing high-traffic road vehicle density, increasing the average velocity of roads, and reducing the arrival time. The best result in the case of London shows a 112.13% increase in average velocity, a 16.47% decrease in density of high traffic roads, and a 61% reduction in arrival time through all the vehicles in the model.

Links and sources

• PDF from original source

Comments