| Popis: |
To address the challenge of conflicting traffic flows that complete on opposing cycle times in a specific phase of the traffic light, we proposed a novel decentralized traffic light control methodology based on the identification of the congestion's bottleneck and its influence. By computing the cost of each congested tree, this methodology, called the Tree Method, prioritizes bottlenecks based on global influence, i.e., their impact on the entire road network. To test this methodology, we used SUMO (Simulation of Urban Mobility) as a traffic simulator to run various simulations with both realistic and abstract Origin-Destination (OD) matrices under different traffic loads. By prioritizing traffic flows based on their global cost, rather than merely local ones, the Tree Method accurately identified the root cause of traffic-tree congestion and its effect along the traffic upstream. It exhibited major improvements in both the throughput and the average travel time measurements. Further in-depth analyses of individual simulation runs and time-based performance demonstrated that the Tree Method is superior compared to other methods, not only in terms of average performance but also for the majority of drivers and throughout most of the time. Another significant advantage of this method is its simplicity of analysis, facilitating real-time adjustments in each cycle and aligning with the inherent feedback dynamics of traffic flows. This feature makes the Tree Method highly suitable for integration into actual traffic control systems. |
