3D-SLAM using iRSP and Marilou Simulator

/in ,

iRSP(intelligent Robot Software Platform) SLAM 3D Execution



The current video demonstrates 3D SLAM capability of a robot using iRSP (Intelligent Robot Software Platform) and anyKode’s Marilou (Robot Simulator).

The map data is updated from the calculated position of the robot using simulated onboard range sensors. Our 3D point cloud library (3D-PCL) optimized for CPU and GPU is being used to process the sensory data real-time.

The robot is driven under 2 drive modes: Automatic, and Non-automatic.

In case of the Non-automatic mode, the user controls the speed and direction of the robot using the control panel.

In case of the Automatic drive mode, the robot is given the position of the final destination, and generates necessary paths as it moves to get to the goal point. If the map is updated and blocks the generated path, the robot immediately re-generates the path not to cross the blocked area.

The specification of the simulated robot is as follows:

1. Motor : 2 DC Motors with incremental encoder
2. 1D range sensor: 8 Sonar sensors
3. 3D range sensor: Xtion-Pro
4. Direction sensor: Absolute Compass Sensor

All the sensors are simulated with the anyKode’s Marilou simulator that is now embedded in the iRSP development IDE as a robot building component.