Automatic Generation of Random Step Environment Models for Gazebo Simulator

Negotiating rough terrain obstacles is a key task of a mobile terrestrial robot. One of the most popular standards in rough terrain modeling is the NIST Random Step Environment (RSE) or Random Stepfield, which allows constructing a broad variety of debris-like environments. This paper proposes a virtual RSE models’ automatic generator LIRS-RSEGen for the Gazebo simulator. We analyzed typical for RSEs obstacles and structures, determined parameters that are required in order to generate a RSE and implemented a generator, which constructs RSE Gazebo worlds that could be further edited or directly used within the Gazebo. Constructed by our generator worlds were validated in the Gazebo using virtual models of TurtleBot3 wheeled robot and Servosila Engineer crawler robot.