The Physics of Life Excellence Cluster of TU Dresden boosts productivity with AI-driven automation

Physics of Life at TU Dresden worked with Wandelbots and the Max Planck Institute to automate repetitive pipetting tasks. A web-based NOVA application lets researchers with limited robotics knowledge instruct a robot to perform work that was previously manual.

Use Case Overview

Repetitive pipetting tasks consume valuable research time and require consistent precision. Yet traditional laboratory automation often demands specialized robotics expertise.Researchers at TU Dresden's Physics of Life cluster developed PipetBot using Wandelbots NOVA, NVIDIA Omniverse, and AI. Through a web application, researchers can instruct a robot using natural language or voice commands, while NOVA handles robot programming and movement planning.The solution significantly simplifies laboratory automation, achieving a fourfold increase in process efficiency and giving researchers more time for scientific work.

ABOUT

Physics of Life is an Excellence Cluster at TU Dresden collaborating with research partners including the Max Planck Institute.

OBJECTIVE

Explore robot automation to remove repetitive pipetting tasks and improve quality in laboratory workflows.

SOLUTION

A web-based application on Wandelbots NOVA uses ChatGPT 4.0 and connected NOVA components to execute robot tasks through written or voice commands.

IMPACT

The project achieved a fourfold increase in process efficiency when handling liquid samples across different test sequences.

Challenges & process

PoL faced productivity delays because pipetting was slow, repetitive, and dependent on researchers performing the work manually. Earlier automation attempts did not fully meet the need for a powerful but easy-to-use solution. The team set up the hardware in NVIDIA Omniverse to validate feasibility, then connected a large language model through a service and API layer. This made it possible to coordinate the pipette tool, camera, robot arm, and NOVA components through a browser-based application. The resulting PipetBot app lets researchers choose pipettes and issue instructions by text or voice while NOVA handles path creation and programming. This helps researchers spend less time on routine handling and more time on scientific work.

“The virtual-first approach in Omniverse made it possible to place the robot flexibly into different laboratory workflows.”

Prof. Dr. Stephan Grill

Professor, Max Planck Institute of Molecular Cell Biology and Genetics

Impact & results

By simplifying workflows and maintaining quality, PoL increased operational efficiency by 400 percent. Researchers can redirect time from routine pipetting to core scientific tasks. The setup also gives the lab more flexibility for future test requirements, making day-to-day operations smoother and easier to adapt as research needs evolve.