Scientific planning of a dedicated network for automated truck cargo transportation is necessary to ensure the transportation system safety and the development of truck cargo automation. Considering the dedicated lanes for automated trucks will reduce the right-of-way of general-purpose vehicles and impact the travel path selection behavior of general-purpose vehicles, firstly, this paper develops a new bi-level programming model for designing a dedicated transportation network for automated trucks to minimize the system travel time by considering general-purpose vehicles' travel path selection behavior from the perspective of the overall system. Then, an improved differential evolutionary algorithm based on a real number coding is proposed to solve the established bi-level program, specifically including a preprocessing strategy based on the problem characteristics and efficient real number coding rules, which not only ensures the feasibility of the solution but also avoids the complex infeasible solution repair process. Finally, the effectiveness of the model and algorithm is verified by the experimental results of the Sioux Falls benchmark network and a large number of randomly generated instances.