Stability and implementation of a cycle-based max pressure controller for signalized traffic networks

Anderson, Leah; Pumir, Thomas; Triantafyllos, Dimitrios; Bayen, Alexandre M.

doi:10.3934/nhm.2018011

Cited by 15 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pumir et al ( 26 ) provide a cyclic extension to MP control with performance gains observed in simulation results. Anderson et al ( 27 ) alter the MP formulation to a cyclic MP controller which performs better than actuated signal controllers. These cyclic MP controllers however require a predefined fixed cycle length, which requires manual fine-tuning based on predicted demand.…”

Section: Literature Reviewmentioning

confidence: 99%

Performance Evaluation of Modified Cyclic Max-Pressure Controlled Intersections in Realistic Corridors

Barman

Levin

2022

Transportation Research Record

View full text Add to dashboard Cite

Max-pressure (MP) signal timing is an actuated decentralized signal control policy. Rigorous mathematical studies have proven the stability properties and established the benefits of different MP policies. However, theoretical studies make assumptions about traffic properties that may not represent reality and whose effects are not explored much in the literature under realistic traffic conditions. This study focuses on examining how different variations of MP perform in realistic scenarios and on finding the most practical policy among those for implementation in real roads. Microsimulation models of seven intersections from two corridors, County Road (CR) 30 and CR 109 from Hennepin County, Minnesota were created. Real-life demand and current signal timing data provided by Hennepin County were used to make the simulations as close to reality as possible. In this paper, we compare the performance of current actuated-coordinated signal control with an acyclic MP and two variations of cyclic MP policies. We present the performance of different control policies as delay, throughput, worst lane delay, and number of phase changes. We also present how different parameters affect the performance of the MP policies. We found that better performance can be achieved with a cyclic MP policy by allowing phase skipping when no vehicles are waiting. Our findings also suggest that most of the claimed performance benefits can still be achieved in real-life traffic conditions even with the simplified assumptions made in the theoretical models. In most cases, MP control policies outperformed current signal control.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Performance Evaluation of Modified Cyclic Max-Pressure Controlled Intersections in Realistic Corridors

Barman

Levin

2022

Transportation Research Record

View full text Add to dashboard Cite

show abstract

“…Decentralized feedback policies for control such as max-pressure [28] address these issues by adjusting phase splits based on differences in upstream and downstream queue lengths. Maxpressure is stable and provides guaranteed bounds for queues and delay under store and forward queuing system [29], but is challenging to implement in real settings due to its hardware and safety constraints [30], [31].…”

Section: A Model-based Coordination Controlmentioning

confidence: 99%

Adaptive Coordination Offsets for Signalized Arterial Intersections using Deep Reinforcement Learning

Diaz¹,

Dailisan²,

Umang³

et al. 2020

Preprint

View full text Add to dashboard Cite

One of the most critical components of an urban transportation system is the coordination of intersections in arterial networks. With the advent of data-driven approaches for traffic control systems, deep reinforcement learning (RL) has gained significant traction in traffic control research. Proposed deep RL solutions to traffic control are designed to directly modify either phase order or timings; such approaches can lead to unfair situations -bypassing low volume links for several cycles -in the name of optimizing traffic flow. To address the issues and feasibility of the present approach, we propose a deep RL framework that dynamically adjusts the offsets based on traffic states and preserves the planned phase timings and order derived from model-based methods. This framework allows us to improve arterial coordination while preserving the notion of fairness for competing streams of traffic in an intersection. Using a validated and calibrated traffic model, we trained the policy of a deep RL agent that aims to reduce travel delays in the network. We evaluated the resulting policy by comparing its performance against the phase offsets obtained by a state-of-the-practice baseline, SYNCHRO. The resulting policy dynamically readjusts phase offsets in response to changes in traffic demand. Simulation results show that the proposed deep RL agent outperformed SYNCHRO on average, effectively reducing delay time by 13.21% in the AM Scenario, 2.42% in the noon scenario, and 6.2% in the PM scenario. Finally, we also show the robustness of our agent to extreme traffic conditions, such as demand surges and localized traffic incidents.

show abstract

“…Several other cyclic algorithms have since been proposed including two by Pumir et al ( 11 ) and Anderson et al ( 12 ) which both had a fixed cycle length. An algorithm by Levin et al ( 13 ) also had a cyclic structure but allowed the cycle length to be lengthened or shortened in response to real-time demand.…”

Section: Introductionmentioning

confidence: 99%

Microsimulation Study Evaluating the Benefits of Cyclic and Non-Cyclic Max-Pressure Control of Signalized Intersections

Robbennolt

Chen

Levin

2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Max-pressure control is a decentralized method of traffic intersection control, making computations at individual intersections simple. In addition, this method of control has been proven to maximize network throughput if any traffic signal control can stabilize the demand. This paper tests max-pressure controllers in a large-scale microsimulation of the downtown Austin network using the microscopic traffic simulation package SUMO. Nine combinations of weight function and method of defining green time are studied to see how different variations on the max-pressure controller compare. It is shown that the way green time is assigned (cyclic or non-cyclic) has a larger impact on performance than the weight function used by the max-pressure controller. Based on these results a new way of assigning green time is devised. This novel controller mirrors the performance of either the cyclic or the non-cyclic controller depending on the geometry and demand. Large-scale simulation shows that this controller compares favorably with existing controllers using metrics of number of waiting vehicles and average travel time. Common problems with non-cyclic control include the higher likelihood of gridlock and the potential for very long waiting times when demand at a single intersection is asymmetric. On the other hand, the cyclic controller is required to allocate green time to every phase even if the demand is low, increasing the loss time. The novel semi-cyclic controller solves these inherent problems with the cyclic and non-cyclic controllers, making it more likely to be implemented by traffic engineers.

show abstract

Stability and implementation of a cycle-based max pressure controller for signalized traffic networks

Cited by 15 publications

References 19 publications

Performance Evaluation of Modified Cyclic Max-Pressure Controlled Intersections in Realistic Corridors

Performance Evaluation of Modified Cyclic Max-Pressure Controlled Intersections in Realistic Corridors

Adaptive Coordination Offsets for Signalized Arterial Intersections using Deep Reinforcement Learning

Microsimulation Study Evaluating the Benefits of Cyclic and Non-Cyclic Max-Pressure Control of Signalized Intersections

Contact Info

Product

Resources

About