RobAR aims to develop an intuitive, user-centric approach using mixed reality (MR). The project enables the operator to interact with the robot's digital twin directly in the physical workspace via an MR headset. The goal is to make the programming process simpler and more efficient.

Instead of robot-centric code adjustments by experts, the MR environment enables the user to define robot trajectories directly on the real workpiece on the shop floor. This seamless workflow includes virtual path planning, simulation and adjustment of trajectories for proactive error correction, and finally, the real execution of the pre-validated program by the physical robot.

Real-time visualization of simulations in the physical world makes it possible to identify collisions and errors early in the planning phase, minimizing setup time on the real system. RobAR thus makes robot programming accessible even for non-experts and contributes to a more efficient and safer paradigm for human-robot interaction.

We are testing the applicability of this method using an industrial welding use case in collaboration with Alpaka Innovation, which demonstrates the workflow in three clearly defined phases, as shown below.