Capacity Estimation of Signalized Intersections Considering Connected Automated Vehicle Observability

With the advancement of sensing, communication, and cooperative capabilities of connected automated vehicles (CAVs), the capacity and operational state of signalized intersections have become increasingly observable and suitable for prospective assessment. However, existing capacity models based on homogeneous traffic assumptions are insufficient to describe the capacity evolution of mixed traffic under varying CAV penetration levels. Motivated by this limitation, this study proposes a quantitative capacity estimation method for signalized intersections considering CAV penetration, serving as an evaluation and prediction baseline for intersection operations. The proposed method improves the CAV gain parameter and accounts for multiple typical car-following states in mixed traffic to derive equivalent headways and spacing coefficients, enabling a continuous estimation of intersection capacity with respect to CAV penetration. Using data from an actual signalized intersection, capacity and saturation trends are analyzed across different movement directions and traffic demand conditions. The results indicate a nonlinear increasing pattern of intersection capacity as CAV penetration rises, with distinct growth rates between straight-through and left-turn movements. The proposed approach provides an engineering-oriented reference for capacity estimation and traffic state prediction under mixed traffic conditions.