Hybrid-phase-enabled multi-mode-band approach to arterial traffic control in mixed traffic environment with self-organized connected and automated vehicles

With the development of connected and automated vehicle (CAV) technology, mixed traffic with CAVs and regular vehicles (RVs) are expected to persist for a long time in the foreseeable future. Research on mixed traffic control often assumes that CAV trajectories can be fully controlled by traffic controllers or that their trajectory planning strategies are known. However, this assumption may not hold in the near term due to limitations in communication technology or concerns over data privacy. In recent years, several studies have addressed traffic control while considering the uncontrollability of CAVs and the limitations of available CAV information. However, these studies typically focus on isolated intersections or the fully CAV environment. This study introduces a hybrid-phase-enabled multi-mode-band-based (HPMM-based) traffic control for arterials with CAV-dedicated lanes in the mixed traffic environment with CAVs and RVs. In this study, CAVs are not controlled by traffic controllers and conduct trajectory planning themselves, which are called self-organized CAVs. For simplicity, they are referred to as CAVs throughout this paper. There are dedicated lanes for CAVs in each arm at each intersection along the arterial. In the proposed model, left-turn and through CAVs share CAV-dedicated lanes and cross the intersection during the shared phases using the standard NEMA ring barrier structure with RVs or during the CAV-dedicated phase. A two-level hierarchical optimization model is developed, which consists of the arterial and the intersection levels. The arterial level introduces a multi-mode-band model to address the signal coordination challenge for arterials with multiple phases (i.e., shared phases and CAV-dedicated phases) and multiple modes (i.e., CAVs and RVs) and CAV-dedicated lanes in the mixed traffic environment. The model is formulated as a mixed integer linear programming problem to maximize weighted bandwidth for CAVs and RVs. At the intersection level, a three-sub-level model optimizes signal timings based on the estimation of RV queue lengths and prediction of CAV passing states without directly controlling CAV trajectories or assuming prior knowledge of their trajectory planning strategies, and a rolling horizon scheme is designed. Numerical results demonstrate the proposed HPMM control framework outperforms existing methods under distinct scenarios: it reduces average vehicle delay and unnecessary stops compared to max-pressure-blue-phase-based control in under-saturated traffic, and surpasses normal control (which lacks CAV-dedicated lane and phase) when CAV penetration rates exceed 10%.