Realistic and Automatic Map Generator for Mobile Robots

摘要

In a lot of sciences as in mobile robots, is always necessary to determine the performance of new algorithms. Profiling is a useful tool during the process of development and testing new algorithms. The use of mobile robot simulators allows to verify and to test the performance and efficiency of new approaches. When an algorithm is evaluated, is very helpful to have a huge testing dataset to ensure, that the roposed methodology is robust for a wide range of possible scenarios. In most of the public dataset, the number of scenarios to probe new algorithms is limited, the maps are hard to modify, and the map description is not standard. Because of that, we introduce a realistic and automatic map generator for mobile robots based on Gazebo due that Gazebo is a very advanced simulator that can be used in conjunction with ROS. This combination enables the easy prototyping and testing of robotics applications, in simulated and in real robots with minimal modifications.