LQ-Informed Rule-Based Variable Speed Limit Control

Variable speed limit (VSL) control is typically implemented using rule-based approaches due to their simplicity, computational efficiency, and ease of deployment using roadside infrastructure. However, these approaches often fail to achieve the full benefits of reducing congestion and improving travel times. Optimal control strategies offer a more effective solution, but rely on infinite-dimensional control laws derived from partial differential equation (PDE) optimal control, which are impractical for direct implementation. This paper presents a novel method to bridge this gap by extracting simplified, implementable rules from the continuous optimal control solution, enabling the creation of an improved rule-based VSL controller. The proposed controller outperforms the baseline controller by reducing average traffic density by 7%, increasing throughput by 46%, and reducing total time spent by 7%, effectively closing the optimality gap by 11%, on a section of Interstate 270 in Columbus, OH.