MuRoSim is a multi robot simulation platform, we implemented what we have learned in virtual reality course and extended them to finally build this system. The objective of the project is to build a platform to test robot mapping, navigation, localization and swarming algorithms in a 3D virtual environment. Basically, the techniques involved in this project are collision detection, physical modeling, and OpenGL graphic rendering. This is only the first version of it and I'm now trying to re program it using some open source libraries such as ODE.

After the system is fired up, you can go to either edit mode or simulation mode. In edit mode, you can use objects such as wall and iRobot to build a map to run the simulation, or you can simply load existing maps. In simulation mode. the main window renders a global view of the arena while the robot cam window renders the vision provided by a virtual camera mounted on top of the selected robot. There is also a robot memory window which shows the real-time map constructed using the sensor data of the robot.

The robot in this project is modeled the same as iRobot Create module which has nonholonomic constraints. In this test version, the only sensors that are modeled are bump sensor and laser range sensor. They are achieved using bounding sphere and ray testing methods. In the demo, we successfully programmed a simple potential field based navigation algorithm to make the robot move to a desired goal autonomously with sensor data.