[Paper Review] Measuring Impact of Adaptive and Cooperative Adaptive Cruise Control on Throughput of Signalized Intersections

Armin Askari, Daniel Albarnaz Farias+2 more|arXiv (Cornell University)|Nov 28, 2016

Traffic control and managementEngineering5 references16 citations

TL;DR

This paper evaluates how Adaptive Cruise Control (ACC) and Cooperative ACC (CACC) improve signalized intersection throughput using three car-following models—Gipps, Helly, and Improved Intelligent Driver Model (IIDM). The IIDM model is found most suitable due to realistic acceleration and deceleration behavior, and CACC significantly boosts throughput only when vehicles are grouped in platoons at bottleneck intersections.

ABSTRACT

To properly assess the impact of (cooperative) adaptive cruise control ACC (CACC), one has to model vehicle dynamics. First of all, one has to choose the car following model, as it determines the vehicle flow as vehicles accelerate from standstill or decelerate because of the obstacle ahead. The other factor significantly affecting the intersection throughput is the maximal vehicle acceleration rate. In this paper, we analyze three car following behaviors: Gipps model, Improved Intelligent Driver Model (IIDM) and Helly model. Gipps model exhibits rather aggressive acceleration behavior. If used for the intersection throughput estimation, this model would lead to overly optimistic results. Helly model is convenient to analyze due to its linear nature, but its deceleration behavior in the presence of obstacles ahead is unrealistically abrupt. Showing the most realistic acceleration and deceleration behavior of the three models, IIDM is suited for ACC/CACC impact evaluation better than the other two. We discuss the influence of the maximal vehicle acceleration rate and presence of different portions of ACC/CACC vehicles on intersection throughput in the context of the three car following models. The analysis is done for two cases: (1) free road downstream of the intersection; and (2) red light at some distance downstream of the intersection. Finally, we introduce the platoon model and evaluate ACC and CACC with platooning in terms of travel time ad network throughput using SUMO simulation of the 4-mile stretch of Colorado Boulevard / Huntington Drive arterial with 13 signalized intersections in Arcadia, Southern California.

Motivation & Objective

• To assess the impact of ACC and CACC on urban intersection throughput using realistic vehicle dynamics models.
• To compare the performance of three car-following models—Gipps, Helly, and IIDM—in estimating intersection capacity and travel time.
• To determine how vehicle penetration rates of ACC and CACC, and platooning behavior, affect traffic flow and queue dynamics.
• To identify optimal conditions for platooning that maximize travel time reduction without creating downstream bottlenecks.
• To provide empirical validation using SUMO simulations on a real-world urban network in North Bethesda, MD.

Proposed method

• Modeling vehicle dynamics using three car-following models: Gipps, Helly, and Improved Intelligent Driver Model (IIDM).
• Simulating traffic flow at signalized intersections under two downstream conditions: free road and red light downstream.
• Analyzing the influence of maximum vehicle acceleration rate and ACC/CACC penetration rates on intersection throughput.
• Introducing a platoon model to evaluate travel time reduction through coordinated vehicle movement.
• Conducting SUMO simulations on a 7-intersection urban network to compare travel times with and without platooning on specific links.
• Applying signal timing and vehicle arrival functions to assess performance under varying traffic compositions and platooning rules.

Experimental results

Research questions

• RQ1 How do different car-following models (Gipps, Helly, IIDM) affect the estimation of intersection throughput under ACC and CACC?
• RQ2 What is the impact of increasing ACC and CACC penetration rates on intersection throughput and travel time?
• RQ3 In what traffic conditions does platooning significantly reduce travel time, and when does it create new bottlenecks?
• RQ4 How does vehicle ordering—grouped at the front vs. interleaved in the queue—affect the performance of ACC and CACC?
• RQ5 Under what conditions does platooning at intersections lead to improved or degraded network performance?

Key findings

• The Gipps model produces overly optimistic throughput estimates due to aggressive acceleration, exceeding theoretical equilibrium flow bounds.
• The Helly model generates unrealistically abrupt deceleration, making it unsuitable for realistic assessment of driver comfort and safety.
• The IIDM model provides the most realistic balance of acceleration and deceleration behavior, making it the best choice for evaluating ACC/CACC impact.
• CACC vehicles achieve significant throughput gains only when grouped in platoons; interleaved CACC and ordinary vehicles perform like pure ACC.
• At 75% CACC penetration, platoons form more frequently and are larger, but this causes oversaturation at downstream intersections, offsetting upstream time savings.
• Enabling platooning on non-bottleneck approaches (e.g., AP3319 and AP3299) provides no travel time benefit and can block cross-street turns, confirming that platooning should be restricted to bottleneck intersections.