Simulace pick and place aplikace s průmyslovým robotem v prostředí Nvidia Isaac Sim

Abstract

This master's thesis deals with the design and implementation of a digital twin of a robotic workplace within the NVIDIA Isaac Sim environment. The theoretical part researches current simulation tools with an emphasis on the NVIDIA Isaac platform and analyzes the physical workplace equipped with a Universal Robots UR10e robot. The practical part serves as a methodological guide describing the creation of the virtual model from CAD data import to physical parameter configuration. The main objective was to implement an automated Pick and Place application. Two approaches to kinematics and trajectory planning were utilized and compared. The first is a native solution within the simulator using a custom extension with the RMPflow algorithm. The second involves external control via the ROS 2 operating system and the MoveIt 2 framework. The conclusion evaluates both methods regarding trajectory reliability, computational complexity, stability and portability to real hardware.