DriveLM
DriveLM copied to clipboard
About the Processing of Trajectory Data
Could you kindly provide a detailed explanation of how the trajectory data is processed? Specifically, could you share more details about the process of discretizing the trajectory data? Thank you for your time and assistance!
Hello thanks for your interest. First we present the step-by-step details of tokenization of the trajectory in supplementary E.2. To be specific in the range of discretization, for history trajectory we set y_range = [-40, 60] and x_range = [-15, 15] (in meters), and for current trajectory we set y_range = [-2.5, 60] and the same for x_range.