The Unimog U5023 is a special truck characterized by an extreme off-road capability and a lot of envelopes like a central tire inflation system, a high number of gears and a shrinkable axle. Because it is possible to drive through rough and extremely cluttered terrain the vehicle is very suited for applied research.
This video is about testing the off-road perception in a lifelike environment that was simulated with the Unreal Engine. The possibility to simulate scenarios creates a good foundation for executing the robot control in the real world. In addition to autonomous robots, various sensors and cameras can be generated. In this case, the robot uses 2 simulated ZED stereo cameras, 2 Intel RealSense D435, 2 SICK TiM 561, 2 SICK LMS 551 laser scanners, an u-blox NEO 7P GNSS, and a Microstrain 3DM-GX5-25 IMU. All sensors correspond to the real hardware layout and have been modeled in the simulation. The U5023 uses the behavior-based control iB2C and REACTiON architecture to safely navigate through the environment.
[ Ссылка ]
[ Ссылка ]
[ Ссылка ]
