Energy saving coordination control of heterogeneous-speed platoons at unsignalized intersections using gap-crossing virtual platoon

Unsignalized intersections are critical bottlenecks for both traffic efficiency and energy consumption in urban networks. Virtual platooning has emerged as a promising coordination paradigm. However, existing approaches coordinate vehicles individually, losing the aerodynamic and scheduling benefits of platoon integrity, and enforce uniform velocity across all turning and through movements, ignoring velocity heterogeneity in real traffic where straight-moving vehicles travel faster while left turning vehicles must slow down for safety and lateral control. Moreover, model conflict zones using fixed geometric shapes or the entire intersection based on multi-lane geometries, lead to conservative estimations of arrival time and distance from individual vehicles to the center of the intersection. To address these issues, a new gap-crossing virtual platoon (GCVP) algorithm was proposed. It enables conflict-free, non-stop coordination of two platoons with distinct native speeds without forcing velocity homogenization through a speed-heterogeneity-tolerant coordination mechanism that preserves each platoon's natural speed while achieving spatiotemporal deconfliction. It improves energy saving and traffic efficiency by reducing unnecessary acceleration and deceleration. To validate the algorithm, a distributed model predictive control (MPC) framework was developed. Compared to First-Come-First-Served (FCFS) policy, the GCVP indicates 10%-25% reduction in average evacuation time and 34.2%-65.7% decrease in fuel consumption across a range of speed scenarios. This paper also discussed the performance of the GCVP algorithm at various time headways. Compared with existing strategies, this GCVP algorithm achieves improvements in average evacuation time by eliminating stop-and-wait delays and reducing speed fluctuations, enabling dynamic speed from 30 to 70 km/h at intersections. It currently supports only two-phase conflict scenarios. However, it presents potential for simultaneously resolute multiple two-phase conflicts, or integration with traffic light systems to enhance traffic efficiency and energy performance at intersections.