Abstract
Various approaches have been employed to model rider behaviors, but replicating complex movements or diverse behaviors requires further consideration. In this study, the aim is to develop a model that can more in detail replicate the behaviors of real riders by leveraging the rapidly evolving technology of deep reinforcement learning. In this initial report, examples are demonstrated of constructing a rider model. Additionally, by utilizing technologies to evaluate riding skills, this study assesses the proficiency of the learning model and attempts to establish a connection between the learning model and reality.
