With the in-depth development of the research on Intelligent Connected Vehicles, there will be a traffic environment of the mixed driving in the future. Therefore, to establish a mixed driving traffic simulation platform to carry out experiments is an important research method. The vehicle agent model and the multi-agent simulation system of mixed traffic flow are constructed based on the theory of multi-agent system, and the simulation platform of a multi-agent system for mixed traffic flow is built by C++. By simulating car-following in continuous traffic flow and comparing simulation data with the data collected from the actual road section, the reliability of the simulation platform is verified. Based on this platform, the simulation experiment of car following in mixed driving traffic flow and human-driven traffic flow were carried out. The comparative analysis results showed that the multi-agent system of mixed traffic flow will run at a smaller space headway and a faster speed. The overall capacity of the road will increase. Meanwhile, the velocity change curve of mixed driving traffic flow speed is relatively smooth, and the stability of traffic flow is better.

CICTP