LIRS-USim: a Gazebo-based Tool for Modeling Urban Environments and Sensory Data Uncertainties

Abstract

Urban Search and Rescue (USAR) robotics deals with emergent situations that occur in urban environments due to natural and human-made disasters. To support early stages of algorithms’ testing and evaluation, we developed an easy to use USAR simulation tool (LIRS-USim) that models typical USAR missions within the Gazebo simulator. The proposed robot operating system based tool is capable of modeling a virtual environment with hazardous zones, constructing a 3D Gazebo world from an arbitrary 2D image and populating it with various obstacles from the Gazebo library, simulating uncertainties and failures of robot’s onboard sensors. To set up a hazardous zone, its location, size and a radiation or chemical contamination shape are defined by a user. Next, any existing in Gazebo robot model with any onboard sensors could be loaded into the 3D world, and probabilities of each sensor uncertainty and failure could be set individually by a user. Moreover, LIRS-USim allows loading several robots of different types into a single Gazebo world and further monitor each robot and each sensor. LIRS-USim was successfully tested with Husky, Warthog, Jackal, and Hector Quadrotor standard robot models. The source code of LIRS-USim is available for free academic use.