This invention discloses a primitive modeling method for mixed vehicle groups in near-
signal zones, comprising the following steps: using roadside sensing equipment to detect vehicles within the sensing area, thereby acquiring acceleration, speed, and position information of connected autonomous vehicles and human-driven vehicles; determining whether a vehicle has entered the near-
signal zone; treating vehicles within the near-
signal zone as nodes in both physical and information spaces; determining the
vehicle type in the near-signal zone; establishing a primitive model of human-driven vehicles in the near-signal zone based on the collected information of vehicles surrounding the current human-driven vehicle; establishing a primitive model of connected autonomous vehicles in the near-signal zone based on the collected information of vehicles surrounding the current connected autonomous vehicle; and establishing a primitive model of mixed vehicle groups in the near-signal zone. This invention can establish a cyber-physical primitive model for multi-lane mixed traffic in near-signal zones, providing theoretical support for alleviating traffic problems in near-signal zones.