With the continuous investment in the field of autonomous truck driving technology, fully autonomous self-driving trucks are set to become a reality soon. As a potential fuel-efficient driving mode of autonomous trucks, the truck platoon has been proposed and attracts great attention to vehicles connecting and control. However, simplified as an extreme long vehicle, autonomous truck platoon will undoubtedly affect the performance of traffic flow as it has an uncertain effect on the behavior of other vehicle drivers such as following and lane changing. Obviously, this effect deserves much attention. Besides, whether to set the special lane for platoons will become a notable issue considering traffic efficiency. To solve these problems, simulation of traffic flow will be the proper way. In this study, a three-lane cellular automaton (CA) model is proposed and the updated forwarding and lane changing rules are described in detail. Truck platoons contain two truck are considered. Based on the CA model, the influence of autonomous truck platoons on traffic flow such as lane changing frequency and average traffic speed is evaluated. Besides, the fundamental diagram of traffic flow is conducted. Compared to the existing cellular automaton models, the lane changing rules adopted in the proposed model differ according to driving scenarios and cell number varies in vehicle types in this study. As the results of numerical simulations show, to limit platoons on the special lane can higher the average speed of traffic flow and reduce the lane changing frequency. The existence of platoons improves the average traffic speed while setting special lane for platoons. However, when the traffic density grows up, this positive effect dissipates gradually.