Based in South Korea, TESOLLO Inc. is redefining robotic automation through precise motion control, advanced embedded systems, and the intelligent integration of AI.
Its flagship product, the DG-5F, is a fully actuated, five-finger anthropomorphic robot hand engineered for human-level dexterity, gripping, and control. Featuring 20 degrees of freedom (DOF), it seamlessly combines high-precision mechanics, optional tactile sensing, and flexible software compatibility, making it ideal for applications in research, industrial automation, and service robots.
Controlling a 20-DoF robot hand like the DG-5F with traditional input methods is challenging. Manual programming or joystick control often leads to delays, low precision, and a steep learning curve. To address this, TESOLLO developed ROS2 packages that make robot hand control more intuitive and efficient.
Beyond showcasing dexterity, the capability of the DG-5F can be expanded to real-world use cases by mounting it on the Franka Research 3 robot arm and pairing it with NOKOV positional tracking. This integration significantly increases the system’s flexibility, mobility, and ability to perform complex manipulation tasks.
With MANUS Metagloves Pro and the MANUS Core ROS 2 Package, a seamless human-to-robot teleoperation pipeline was developed using TESOLLO DG-5F. The NOKOV optical tracking system serves as the positional anchor for the Franka Research 3 robot arm, providing spatial reference for precise alignment between human and robot movements.
Through this collaboration, TESOLLO demonstrated how the combination of MANUS Metagloves Pro, ROS2 integration, and optical tracking can enable highly dexterous teleoperation workflows.
However, fine-grained teleoperation still faces challenges due to the kinematic differences between human and robot hands. MANUS’ endpoint sensor streaming offers a path forward, allowing robot hands like the DG-5F to learn and refine their motion models over time.
This project highlights how combining high-fidelity motion data with adaptive control can pave the way toward more intelligent, data-driven manipulation systems in the future.
