Dissipation of stop-and-go waves via control of autonomous vehicles: Field experiments (1705.01693v1)

Abstract: Traffic waves are phenomena that emerge when the vehicular density exceeds a critical threshold. Considering the presence of increasingly automated vehicles in the traffic stream, a number of research activities have focused on the influence of automated vehicles on the bulk traffic flow. In the present article, we demonstrate experimentally that intelligent control of an autonomous vehicle is able to dampen stop-and-go waves that can arise even in the absence of geometric or lane changing triggers. Precisely, our experiments on a circular track with more than 20 vehicles show that traffic waves emerge consistently, and that they can be dampened by controlling the velocity of a single vehicle in the flow. We compare metrics for velocity, braking events, and fuel economy across experiments. These experimental findings suggest a paradigm shift in traffic management: flow control will be possible via a few mobile actuators (less than 5%) long before a majority of vehicles have autonomous capabilities.

• The paper demonstrates that controlling a minimal number of AVs can reduce velocity fluctuations by up to 80.8%, stabilizing traffic flow.
• It details the use of FollowerStopper and PI controllers to smooth acceleration and deceleration, mitigating stop-and-go instabilities.
• Experiments revealed a 42.5% improvement in fuel economy and increased traffic throughput, highlighting practical traffic management benefits.

Dissipation of Stop-and-Go Waves via Control of Autonomous Vehicles: Field Experiments

This paper explores the potential of autonomous vehicles (AVs) to mitigate stop-and-go traffic waves without the need for extensive infrastructural changes. The authors demonstrate through field experiments that controlling a minimal number of AVs in a traffic stream can significantly reduce traffic instabilities, offering a promising perspective for future traffic management.

Experimental Methodology

The experiments employ a single-lane circular track, with key phases including initial setup, driving under human control, wave-generation, and finally, transitioning to autonomous control. The experiments differ primarily in the control strategy of a single vehicle within the flow, either through an automated system or by a human driver trained to follow specific velocity targets.

Control Strategies

1. FollowerStopper Controller: This strategy involves an AV maintaining a set target speed unless safety constraints necessitate adjustments based on measured gaps and relative vehicle velocities. The aim is to smooth the driving profile without abrupt accelerations or decelerations.
2. PI Controller with Saturation: Utilizing an average speed derived from real-time measurements, this controller adjusts the AV's speed dynamically, aiming to minimize unnecessary braking and acceleration events while maintaining a safe following distance.
3. Human Driver Implementation: A human operator mimics the control strategy used by automated systems, delivering a practical demonstration of the effects when advanced driving training is employed.

Key Results

• Velocity Standard Deviation: Across experiments, the application of control significantly reduced velocity fluctuations, with reductions up to 80.8% in certain scenarios, implying enhanced flow stability.
• Fuel Economy: There was a marked improvement in energy efficiency, achieving up to 42.5% reduction in fuel consumption, especially when AVs operated under optimal control conditions.
• Reduction of Braking Events: The rate of excessive braking events dropped substantially, indicating a smoother flow and reduced likelihood of wave propagation.
• Traffic Throughput: The throughput was increased in certain experiments, highlighting the potential of AVs to enhance traffic efficiency.

Theoretical and Practical Implications

The findings underscore a shift from traditional traffic management strategies, which rely heavily on static infrastructure, to dynamic, vehicle-based control methods. The paper suggests that even a minority of intelligently controlled AVs within traffic can substantially impact overall flow dynamics, effectively dissipating congestion waves and reducing fuel usage and emissions.

Future Directions

Research is likely to focus on:

• Extending these findings to multilane roads and urban networks.
• Exploring integration with connected vehicle technologies to coordinate AV actions more effectively.
• Investigating the impact of varying penetration rates of AVs on flow dynamics.

By harnessing the capabilities of AVs, this paper provides a foundation for future traffic systems that are efficient, economical, and environmentally friendly. The potential applications span beyond roadways, offering insights into controlling other distributed systems characterized by human-influenced dynamic behaviors.