Share Knowledge: Open Source Code

As a research institution for technology transfer, the FZI shares knowledge by developing open source code within the project requirements. An overview about the publication of Open Source Code that emerged within the framework of FZI projects can be found here.

Please find the license terms in the respective description at GitHub.


The GPU-Voxels library allows it to efficiently recognise collisions between geometrical models and 3D sensor data. Collision detection is the basis for two essential problems in robotics: The planning of robot movements as well as the monitoring of the performance of movements within dynamic environments.

Thus, the developed library contains algorithms and data structures that are designed for highly parallel processing via graphics processors: A load balancing octree, voxel maps and voxel lists as well as related algorithms combine these data structures. Furthermore, a shared memory visualisation is included to monitor all data in real time.

Open Source SVH Hand Driver

By order of the company SCHUNK GmbH & Co. KG, an open source ROS driver for the five-fingered hand SVH was developed at the FZI.

Besides the low level driver, the package also included a comfortable high level control interface for the hand. Furthermore, a detailed 3D model for grip planning, initialising routines as well as extensive possibilities for parametrisation of the hardware controllers are included.

The package will be maintained by the FZI and refined with the help of the ROS community.


Open Source AAL Middleware

The universAAL framework is a reference architecture and implementation for AAL middlewares. Numerous methods and tools support the different roles involved in the development, installation, configuration and maintenance of AAL products. The reference implementation of AAL middleware was tested in eight Living Labs in Europe and installed in 7000 households within the framework of the project ReAAL (