Confi-GRIP

Software configurator for robotic grippers

  • Artificial Intelligence
  • Robotics
  • Synthetic Data
  • Simulation
  • Machine Learning
  • 3D Visualization
  • Application Development
  • Frontend
  • Backend

At a Glance:
Project Metrics

Period:May 2026 - April 2029
Budget:BMFTR-funded · funding code 16IS26008B
Team:Collaborative project with four partners
Industry:Special-purpose machinery, robotics and industrial automation
Area of Application:Flexible assembly and order picking
Technologies
Python is an interpreted high-level programming language for general purposes. Known for simple syntax, promotes readable code and reduces maintenance costs.PyTorch is a popular open-source deep learning framework with a dynamic computation graph, ideal for research and production.ManiSkill is an open-source simulation platform for training and evaluating robotic tasks.Three.js is a JavaScript library for creating 3D graphics in web browsers using WebGL. Simplifies complex 3D rendering for interactive web experiences.Vue is a JavaScript library for building complex, dynamic web applications easily. Vue.js is a newer technology with active development and improvements.FastAPI is a tool for creating APIs. APIs allow data exchange between software components.An open-source tool that simplifies application deployment through container virtualization.

The project

New components regularly present special-purpose machinery builders with the same challenge: finding, designing and testing a suitable gripper and a reliable gripping strategy for the real machine. Today, this work combines extensive experience with time-consuming trial and error.

The ongoing collaborative project Confi-GRIP is developing a web-based software configurator to address this challenge. It is intended to derive suitable gripping strategies and gripper designs from component CAD data and requirements, test them in physics-based simulations and make several solutions comparable. The long-term goal is to reduce the design effort for new assembly and order-picking tasks from weeks to hours.

Confi-GRIP is a research and development project. The platform, models and demonstrators are being developed and validated together with the consortium partners throughout the project.

Our contribution

As a consortium partner, Helm & Walter develops the AI and software components that identify the most promising candidates for the computationally expensive physics simulations. We combine machine learning with a user-oriented platform for design and simulation.

AI core components

We develop surrogate models and other machine-learning methods to select promising gripping and design candidates faster and more reliably.

Interfaces and interaction

We are responsible for the APIs and user interfaces controlling grippers and simulations. This includes parameterisation, user guidance and UI/UX for collaboration between people, models and simulations.

Data and platform

We develop data management, the delivery and visualisation of simulation data, and SaaS functionality for the overall platform.

From component to informed decision

The configurator links four steps: CAD and requirement inputs, AI-supported preselection, physics-based evaluation and a comparable presentation of the results. The computationally expensive simulations can therefore focus on candidates that are particularly promising for the specific application.

The consortium

Confi-GRIP is being developed by four partners: Hiersemann Prozessautomation GmbH (coordinator), LSA GmbH, Helm & Walter IT-Solutions GmbH and the Fraunhofer Institute for Machine Tools and Forming Technology IWU. Together, they combine expertise in special-purpose machinery, automation, software engineering and scientific simulation.

Funding

Funded by: the Federal Ministry of Research, Technology and Space (BMFTR) under the KMU-innovativ: Information and Communication Technologies (ICT) programme.

Funding code for Helm & Walter: 16IS26008B · Funding period: 1 May 2026 to 30 April 2029

The work is carried out as a collaborative research project. The consortium will test the development results with demonstrators and real-world application scenarios.

BMFTR funding logo

Ready to take the first step?

Facing a software challenge? We are the experts who pick up where others leave off! Let's work together to optimize your existing software or start a custom project. Contact us for a non-binding initial analysis. Your project deserves a solution with foresight.

How can we assist you?