Abstract: Stop-and-go traffic waves, often called "phantom jams," significantly reduce freeway capacity and increase safety risks. This review bridges the gap between two historically isolated control strategies: infrastructure-based Variable Speed Limits (VSL) and vehicle-based Jam-Absorption Driving (JAD). By synthesizing decades of research, the authors identify critical theoretical mismatches and the overlooked challenge of "secondary waves" caused by control interventions, providing a roadmap for integrated traffic management using connected and automated vehicles.
What Problem Does This Paper Address?
Freeway drivers frequently encounter sudden standstills that clear without an apparent cause. These stop-and-go waves propagate upstream, reducing road capacity by up to 25% and spiking accident risks. To combat this, engineers have developed two primary tools:
- Variable Speed Limits (VSL): Utilizing electronic gantries to slow incoming traffic before it reaches a bottleneck.
- Jam-Absorption Driving (JAD): Employing "pace car" maneuvers (slow-in, fast-out) by individual vehicles to create a buffer that absorbs the wave.
Despite their shared goal, researchers in these two camps rarely collaborate. VSL studies are grounded in macroscopic flow theory but require expensive infrastructure, while JAD studies offer flexibility but often lack rigorous field testing and theoretical depth. This paper addresses this "disconnection" to prevent redundant research and foster a unified control framework.
Evolution and Methodology
The authors categorize the evolution of both strategies to highlight their diverging paths. VSL has moved from early Model Predictive Control (MPC) to Kinematic Wave Theory (KWT) approaches like the "SPECIALIST" algorithm. Conversely, JAD has evolved from simple car-following rules to modern AI-driven strategies that leverage Connected and Automated Vehicles (CAVs) as mobile traffic dampers.
To evaluate these fields, the authors use a comparative assessment across dimensions such as robustness to randomness, field test maturity, and generalizability. They highlight standard datasets like NGSIM and highD, as well as simulation environments like VISSIM and SUMO, which serve as the primary proving grounds for these algorithms.
Key Insights and the "Secondary Wave" Trap
A major finding of the review is the "Secondary Wave" Trap. Both VSL and JAD can inadvertently trigger new traffic jams behind the control point if traffic is slowed too abruptly. This phenomenon is frequently overlooked in theoretical models but is a critical hurdle for real-world deployment.
Furthermore, there is a Fundamental Diagram (FD) Mismatch. VSL research typically relies on "capacity drop" theory, while JAD research often utilizes simpler triangular models. Bridging these mathematical assumptions is essential for a unified theory of traffic flow control.
Key Results and Maturity
- VSL Success: Field tests, such as the Dutch A12 experiment, have proven VSL can resolve waves, though it tends to "over-activate" in response to minor oscillations.
- JAD Efficiency: Simulations indicate that even a single "absorber" vehicle (a penetration rate of <1%) can significantly smooth traffic flow if positioned correctly.
- The Maturity Gap: VSL is theoretically mature but inflexible; JAD is highly flexible but lacks the stochastic (randomness-based) modeling required for reliable deployment in human-driven environments.
Future Directions
The authors advocate for a "plug-in" modeling structure where traffic managers can choose between VSL (infrastructure-heavy) or JAD (vehicle-centric) based on available resources. Future work must prioritize large-scale JAD field tests and incorporate the "chaos" of human behavior and weather into currently deterministic models to bridge the "Sim-to-Real" gap.
Summary updated